tag:www.econsultancy.com,2008:/topics/sentiment-analysis Latest Sentiment analysis content from Econsultancy 2016-05-09T15:15:00+01:00 tag:www.econsultancy.com,2008:BlogPost/67759 2016-05-09T15:15:00+01:00 2016-05-09T15:15:00+01:00 Expanding your marketing playbook with predictive analytics John Kelly <p>For some people, their feelings about those horses and what they say about the brand transcends more logical price or taste comparisons.</p> <p>While categories with basic emotional appeals may always exist, today’s customer is more sophisticated and the marketing necessarily more complex.</p> <p><iframe src="https://www.youtube.com/embed/dlNO2trC-mk?wmode=transparent" width="560" height="315"></iframe></p> <p>Simultaneously, there is great pressure on chief marketing officers to deliver organic growth for their companies. The playbook must have more pages!</p> <p>One of those new pages is <a href="https://econsultancy.com/blog/64743-predictive-analytics-machine-learning-and-the-future-of-personalization/">predictive analytics</a>, which empowers marketers by providing insights into customer behavior and how certain strategic decisions can increase sales.</p> <p>Those who can best predict the customer and act on those insights will ultimately take market share from their less attuned competition.</p> <p>In fact, <a href="https://econsultancy.com/blog/67315-eight-tips-for-getting-corporate-buy-in-for-data-analytics/" target="_blank">83% of marketers</a> say they now use predictive metrics to build competitive experiences and make smart product decisions.</p> <h3>Predictive Analytics X's &amp; O’s</h3> <p>Predictive analysis of data allows you to play out different “what if” scenarios so you can develop campaigns that achieve optimal growth.</p> <p>Getting more specific, here are four predictive analytics plays CMOs might consider, depending on their business:</p> <h4>1. Sentiment analysis</h4> <p>A sentiment analysis identifies and categorizes the opinions expressed in a section of text to determine whether the writer’s attitude is positive, negative, or neutral.</p> <p>If customers post product reviews on their blogs or discuss your services on social media, a sentiment analysis will dissect their text for clues about their satisfaction levels.</p> <p>Scoring the relative sentiment expressed in mediums from social media commentary to call center transcripts, and comparing peaks and valleys and their drivers over time is powerful.</p> <p>It essentially transforms the idea of “<a href="https://econsultancy.com/reports/voice-of-the-customer-listen-measure-act/">voice of the customer</a>” from a concept to a real tool to sharpen campaigns, products, and customer service operations. </p> <h4>2. Hedonic analysis</h4> <p>Hedonism means the pursuit of self-indulgence, so it follows that a hedonic analysis studies consumers’ preferred features.</p> <p>Instead of guessing what people want, you identify which options are most attractive to them. If you’re an automotive manufacturer, your next model could include dozens of feature combinations.</p> <p>Understanding which combinations will achieve the highest value in the marketplace is useful for optimizing product offerings to best match customer desires and, ultimately, demand. </p> <h4>3. Credit analysis</h4> <p>Credit rating agencies originated in the early 1900s, making credit the oldest form of predictive analysis and a basic tenet of modern businesses.</p> <p>Whenever you offer financing, you’re asking, “Does the customer have both the good faith and the ability to pay?”</p> <p>If you can tweak your approval formulas to qualify more candidates without triggering higher default rates, you’ve mastered organic growth creation. Someone figured that out already, right?</p> <p>The difference today is the availability of abundant supplemental behavioral data that, it turns out, have a lot to do with understanding whether the customer will be a faithful creditor.</p> <p>Social media patterns, cellphone ownership, and usage are just a few ways credit can be predicted more reliably, which is very useful in markets that don't have traditional credit-prediction methods like FICO scores.</p> <h4>4. Churn propensity analysis</h4> <p>Another behavior prediction measurement, churn propensity analysis, anticipates how likely customers are to cancel their annuity contracts.</p> <p>This metric is perfect for <a href="https://econsultancy.com/blog/66034-the-pros-and-cons-of-subscription-ecommerce-models/" target="_blank">subscription-based businesses</a> as well as credit card providers, media brands, cellphone carriers, and direct response companies.</p> <p>If you know which customers will leave and when, then you can court them to remain with you by employing preventive marketing measures.</p> <p>Altered service options and tailored discounts help you avoid losing their business long-term.</p> <p>One rule of marketing that hasn’t changed is that it’s less expensive to <a href="https://econsultancy.com/blog/11051-21-ways-online-retailers-can-improve-customer-retention-rates/">retain a customer</a> than to acquire a new one.</p> <h3>Use with caution</h3> <p>Be careful with the possibilities enabled with the new playbook.</p> <p>Predictive analytics enable you to engineer “wins” such as mastering <a href="https://econsultancy.com/blog/65327-why-dynamic-pricing-is-a-must-for-ecommerce-retailers/">dynamic pricing</a>, increasing profits, and retaining more customers, but those victories aren’t achieved by numbers alone.</p> <p>The most successful CMOs marry in-depth metrics with original thinking to deliver standout campaigns. </p> <p>“Data analytics — big data — is not a substitute for innovation. It’s not a substitute for creative thinking and leadership,” said Sandeep Sacheti, executive vice president for customer information management and operational excellence at Wolters Kluwer.</p> <p>CMOs who can tap into both the logical and creative will achieve greater professional success today and in the future.</p> <p>Anthony Scriffignano, senior vice president and chief data scientists at Dun &amp; Bradstreet, echoed this idea: “The environment is busy and chaotic. Marketers are under a lot of stress and pressure to deliver the growth,” he said.</p> <blockquote> <p>This new data era moves away from the creative and the gray to the very specific, the black and white, the binary. They’re not going to deliver that growth with data-driven analytics unless they take the careful time and process to do it right.</p> </blockquote> <h3>Know your data strategy</h3> <p>With some success, perhaps in the form of a pilot project, you’ll want to get much deeper.</p> <p>Consider your overall scheme if you and your company are going to be real data players.</p> <h4><strong>1. Get intimate with your data</strong></h4> <p>Verse yourself in your analytics and look for gaps in the numbers.</p> <p>It’s worth identifying your current analytics state and your ideal. What would the perfect setup look like, and which metrics do you need to get there?</p> <p>But be discerning about which numbers you emphasize. You may not need every metric available to you, and your vision for how to use your data should evolve as your company grows.</p> <p>Blind pursuit of an ideal state that doesn’t match your organization is wasteful, so evaluate which information moves you toward your goals.</p> <p>As you determine the most valuable threads, you’ll have to make spending decisions about how to gather, validate, integrate, and analyze the right data.</p> <h4><strong>2. Think scientifically</strong></h4> <p>You don’t need to be <a href="https://econsultancy.com/blog/67203-data-analysts-vs-data-scientists-what-s-the-difference/">a data scientist</a> yourself to use analytics effectively, but you do need to hire some.</p> <p>Point-and-click technology is not an honest replacement for the hard work data scientists perform to root out causal relationships, and it creates a false sense of control over the numbers.</p> <p>Your competitors are already catching on to the importance of data scientists and going after top talent via direct hires or consultancies.</p> <p>You can’t afford to fall behind in this area, and data science is not a DIY game. Recruit the best data team you can find by offering candidates big, bold projects and a culture that values their work. </p> <h4><strong>3. Know what matters to you</strong></h4> <p>Do you know your five most impactful marketing data metrics? I don’t mean in concept; I’m talking about hard numbers. Have you tested your data, assessed its weaknesses, and identified statistically proven causal relationships?</p> <p>For instance, you might find that if you drop service wait times by 10%, you can increase prices by 5%.</p> <p>The data might indicate that changing a single term of your offering like the subscription cancellation policy and its accessibility could reduce customer churn by 10%.</p> <p>Make a plan to discover the undeniable facts of your business performance. As a leader or marketer, you must deal in proven, impactful metrics, or what I call the lowest-hanging analytical fruit. </p> <p>Predictive data and <a href="https://econsultancy.com/events/future-of-digital-marketing-london/">the future of marketing</a> are intertwined. “Data analytics is going to be standard, expected practice in every business function and business model,” Sacheti said.</p> <blockquote> <p>Just like HR is a function that exists everywhere in every company and finance exists everywhere in a company, big data scientist is going to be a profession that’ll be expected in every company.</p> </blockquote> tag:www.econsultancy.com,2008:BlogPost/67621 2016-03-08T15:21:02+00:00 2016-03-08T15:21:02+00:00 Is Twitter getting too noisy for high-profile users? Patricio Robles <p>When asked about his declining activity levels, Curry gave Barr a straightforward explanation...</p> <blockquote> <p>...we started having conversations about it and I asked why it was and his response was "it’s just too much of a hassle, it’s too much noise, there’s too much friction to make it worthwhile to spend time doing this."</p> </blockquote> <p>So Barr, who previously worked at Nike, teamed up with Curry and a former Nike colleague, Jason Mayden, to launch Slyce, a new social platform that the trio hopes will provide high-profile figures like Curry with a better way to interact with fans. </p> <blockquote class="twitter-tweet"> <p lang="en" dir="ltr">Join me on <a href="https://twitter.com/hashtag/Slyce?src=hash">#Slyce</a> for a live Q&amp;A on What a day....... <a href="https://twitter.com/hashtag/TeamSlyce?src=hash">#TeamSlyce</a> <a href="https://t.co/SdLj8yStza">https://t.co/SdLj8yStza</a> <a href="https://t.co/Y6jqC9dTc7">pic.twitter.com/Y6jqC9dTc7</a></p> — Stephen Curry (@StephenCurry30) <a href="https://twitter.com/StephenCurry30/status/706718278802997250">March 7, 2016</a> </blockquote> <p>Last month, Curry used his Twitter account, which has over 4.5m followers, to promote a live Q&amp;A session held on Slyce, and the startup has recruited a handful of other athletes to use the service.</p> <p><a href="http://ftw.usatoday.com/2016/03/steph-curry-slyce-davidson">According to</a> Curry, "It’s just kind of having control of your own voice." His co-founder, Barr, elaborated...</p> <blockquote> <p>(Traditional) communication channels have become extremely one sided. And then you have a guy like Steph who has millions of followers the only way he can really interact with them is posting a picture and then walking away from the platform.</p> </blockquote> <p>Slyce aims to faciliate better, more meaningful interactions with fans.</p> <p>When Curry held his Q&amp;A session, the company sorted through the questions and selected those that were of the highest quality.</p> <p>Slyce says that selection process is 90% manual today but it hopes to be able to add greater automation as it gathers more data it can analyze.</p> <h3>A new trend?</h3> <p>It's no surprise that athletes like Curry are investing in social media. And there's plenty of reason to believe that some of the investments they're making could pay off.</p> <p>For example, celebrity-backed Shots, which counts singer Justin Bieber and boxer Floyd Mayweather as investors and early-adopters, was at one point <a href="http://www.cnbc.com/2014/11/25/twitter-in-talks-with-bieber-backed-app-shots-source.html">reported</a> to be in acquisition talks with Twitter.</p> <p><strong>But Slyce isn't just a new social media app. </strong>It's a different kind of social platform that aims to facilitate a different kind of dialog between high-profile users like Curry and their followers.</p> <p>Can such a platform find success? While celebrity participation will almost certainly drive some level of adoption, it's still doubtful that celebrities will be able to direct enough of the online conversation to platforms they control that they can ignore platforms like Twitter, Facebook <a href="http://www.bloomberg.com/features/2016-how-snapchat-built-a-business/">and Snapchat</a>.</p> <p>Indeed, the fact that Curry used Twitter to promote his Slyce Q&amp;A highlights just how valuable his Twitter following is.</p> <p>But if Curry doesn't stay active enough on Twitter or he uses his Twitter account primarily as a vehicle to direct followers to other platforms, the value of that following could decrease.</p> <p>For this reason, it's important that high-profile users ensure they're using all of the solutions at their disposal, such as <a href="https://econsultancy.com/blog/64864-16-essential-social-media-management-tools-we-love">social media monitoring tools</a>, which can help them separate the wheat from the chaff on platforms like Twitter.</p> <p>Fortunately, recognizing just how important high-profile users are to their audiences, the platforms themselves are also looking for ways to better connect celebrities and fans.</p> <p>For instance, <a href="https://www.facebook.com/about/mentions/">Facebook Mentions</a> aims to provide "a better way for actors, athletes, journalists and other public figures to stay in touch with their followers and the people and topics they care about."</p> <p>While there's no reason to believe that there isn't room for new platforms like Slyce, celebrities will still need to actively participate on the largest platforms if they want to be a part of the conversation.</p> tag:www.econsultancy.com,2008:BlogPost/67573 2016-02-24T16:48:49+00:00 2016-02-24T16:48:49+00:00 Facebook rolls out Reactions, an extension to the Like button Patricio Robles <p>As Facebook product manager Sammi Krug <a href="http://newsroom.fb.com/news/2016/02/reactions-now-available-globally/">explained</a>:</p> <blockquote> <p>For more than a year we have been conducting global research including focus groups and surveys to determine what types of reactions people would want to use most.</p> <p>We also looked at how people are already commenting on posts and the top stickers and emoticons as signals for the types of reactions people are already using to determine which reactions to offer.</p> </blockquote> <p>Ultimately, Facebok settled on six emotions: Like, Love, Haha, Wow, Sad or Angry.</p> <p><img src="https://assets.econsultancy.com/images/0007/2322/facebookreactions.gif" alt="" width="396" height="114"></p> <p>Users are not required to use Reactions. They can still opt to simply Like content, and the ubiquitous Like button will remain in place.</p> <p>To provide a Reaction, users can hover over the Like button in the desktop experience or hold down the Like button in mobile experiences to expose the additional Reactions.</p> <p><iframe src="https://player.vimeo.com/video/156501944?title=0&amp;byline=0&amp;portrait=0" width="500" height="281"></iframe></p> <h3>What Reactions mean for the News Feed</h3> <p>Facebook's Krug <a href="http://newsroom.fb.com/news/2016/02/news-feed-fyi-what-the-reactions-launch-means-for-news-feed/">says</a> that all Reactions will be treated equally by its algorithm as far as the News Feed is concerned:</p> <blockquote> <p>Initially, just as we do when someone likes a post, if someone uses a Reaction, we will infer they want to see more of that type of post.</p> <p>In the beginning, it won’t matter if someone likes, “wows” or “sads” a post — we will initially use any Reaction similar to a Like to infer that you want to see more of that type of content.</p> </blockquote> <p>But over time, Facebook will evaluate how different Reactions are used and could tweak its algorithm to weigh Reactions differently.</p> <p>If this happens, it's not inconceivable that Reactions like Love, Haha and Wow could be more favourable than Like, Sad or Angry, so brands will want to pay attention to developments in this area to ensure they can incorporate Reaction-related algorithm changes into their Facebook content marketing strategies.</p> <h3>Brands could benefit</h3> <p>While there's risk that Reactions could some day complicate life for brands as far as the News Feed algorithm is concerned, they also stand to benefit from users' newfound ability to express emotion when reacting to content.</p> <p>Right now, much of the sentiment analysis that brands perform on social platforms like Facebook is qualitative and based on free-form unstructured data, like comments.</p> <p><strong>With Reactions, brands will be able to gauge sentiment on the world's largest social network in a quantitative way. </strong></p> <p>The importance of this shouldn't be underestimated, and savvy brands will look to analyze Reactions, which can be monitored under Page Insights, as quickly as possible.</p> tag:www.econsultancy.com,2008:BlogPost/67499 2016-02-09T15:13:00+00:00 2016-02-09T15:13:00+00:00 Quicken Loans shows the perils of Super Bowl ads Patricio Robles <p>But a super-expensive Super Bowl ad doesn't guarantee advertisers the results they hope for.</p> <p>Mortgage lender Quicken Loans learned that lesson the other day when its Super Bowl ad sparked jeers on social media.</p> <p>The ad promoted the company's Rocket Mortgage tool, which aims to simplify and speed up the process of applying for and obtaining a home loan.</p> <p><iframe src="https://www.youtube.com/embed/QlRm6Y5iVfw?wmode=transparent" width="560" height="315"></iframe></p> <p>But some viewers drew a parallel between the tool and the loose mortgage lending standards that caused the 2008 financial crisis, and took to social platforms like Twitter to criticize the company.</p> <blockquote class="twitter-tweet"> <p lang="en" dir="ltr">Are we really offering mortgages with almost no underwriting again? Was that the <a href="https://twitter.com/hashtag/RocketMortgage?src=hash">#RocketMortgage</a> message? <a href="https://twitter.com/hashtag/SuperBowl?src=hash">#SuperBowl</a></p> — Pat Kiernan (@patkiernan) <a href="https://twitter.com/patkiernan/status/696491501237436417">February 8, 2016</a> </blockquote> <p>Quicken Loans responded quickly to dispel the concerns: Rocket Mortgage makes the mortgage process quicker and easier for consumers but the loans are still subject to stringent underwriting standards.</p> <p>But even responding to negative buzz was turned into fodder for more negative buzz.</p> <blockquote class="twitter-tweet"> <p lang="en" dir="ltr">The only thing more predatory than <a href="https://twitter.com/hashtag/RocketMortgage?src=hash">#RocketMortgage</a> lending is the aggressive PR team responding to every negative tweet</p> — Jo Kimbrell (@jokimbrell) <a href="https://twitter.com/jokimbrell/status/696727998498795520">February 8, 2016</a> </blockquote> <p>Clearly, Quicken Loans' costly Super Bowl ad missed the mark with some consumers.</p> <p>No company spending $5m on an ad during the Super Bowl would want to find itself in a situation where it's having to defend itself publicly on a day when it should be basking in Super Bowl glory.</p> <h3>Traffic tells another story</h3> <p>Social buzz is just one part of the story, however, and negative sentiment can be deceptive.</p> <p>Small minorities of vocal users can create a firestorm that appears larger than it really is.</p> <p>With that in mind, Quicken Loans says that despite the criticism, its Super Bowl ad was a success. According to Jay Farner, the company's CMO, the ad drove 14,000 people to the Rocket Mortgage site within the first minute its ad was broadcast.</p> <p>"The win was driving folks to the site," <a href="http://adage.com/article/special-report-super-bowl/quicken-loans-responding-backlash-super-bowl-ad/302585/">he told</a> AdAge.</p> <p>14,000 visitors in a minute does seem like a win, but is it?</p> <p>Even if one assumes that Quicken Loan's Super Bowl was responsible for driving 100 times that amount of traffic – 1.4m visits – it's not clear just how big a win the ad really delivered.</p> <p>While under this completely hypothetical scenario a cost of roughly $3.50 per visitor might compare very favorably to typical CPCs for mortgage search terms on Google AdWords, ultimately cost per acquisition is the metric that matters most and there are logical reasons to believe that Super Bowl ads probably don't deliver high conversions.</p> <p>After all, many visitors are likely visiting without strong intent.</p> <p>Of course, most Super Bowl advertisers aren't concerned solely with action. The Super Bowl is the quintessential branding exercise, so there are intangible considerations that can't fully be quantified.</p> <p>But even so, Quicken Loans' experience demonstrates that despite their high costs, high-profile ad slots can still carry higher-than-anticipated risks and future Super Bowl advertisers will want to take those into consideration when crafting their ads.</p> <p>And some might even find it wise to consider <a href="https://econsultancy.com/blog/67488-four-tips-for-capitalizing-on-the-super-bowl-without-spending-millions">trying to capitalize on the event with lower risk, lower cost strategies</a>.</p> tag:www.econsultancy.com,2008:BlogPost/67137 2015-11-02T14:49:00+00:00 2015-11-02T14:49:00+00:00 Social monitoring & listening: What is it and do you need it? Evan Dunn <p>But once you understand the underlying framework behind each of these tools - data collection, normalization and natural language processing - it is surprisingly easy to distinguish between the value of different tools.</p> <p>And from there, to embed them into strategies that drive marketing organizations further through research and analytics.</p> <h3>Social listening as a tool for data-driven growth</h3> <p>The most common question I get when talking with people about social listening is “What do I do with this data?”</p> <p>The answer is:</p> <ul> <li>Understand your audience.</li> <li>Care for your audience.</li> <li>Activate your audience.</li> <li>Benchmark against your competition.</li> <li>Optimize your campaigns.</li> </ul> <p>Research is, after all, the beginning of better marketing. </p> <p>Yes, it takes a bit of creativity to apply social listening data to real business problems. That’s why I’m writing a blog right now.</p> <h3>First, a definition:</h3> <p>Social listening, sometimes called social monitoring, refers to technology that has the following properties:</p> <ul> <li>Scans major social networks for instances of a keyword or key phrase.</li> <li>Scans other sites - blogs, forums, news, etc - for instances of a keyword or key phrase.</li> <li>Analyzes the results of these scans with Natural Language Processing (NLP) for sentiment.</li> <li>Analyzes the results of these scans for various actionable insights, including major topics and key influencers.</li> </ul> <p>Capabilities vary widely among tools. Many social listening platforms are not worth the money they charge, others are reasonably priced.</p> <p>Prices, by the way, range between $10 and $10,000 per month - or higher for custom and volume-based pricing.</p> <p>We will be discussing the ‘Enterprise Social Listening’ tools on the higher end, between $1,000 and $10,000 per month.</p> <h3>Understand your audience: Market research &amp; product research with social listening</h3> <p>Most social listening tools scan hundreds of millions of websites for mentions of certain keywords.</p> <p>This means that the results you’re looking at are really a sample set of the entire 'social web'.</p> <p>So if you want to know what people online think about 'smartphones' - or if you’re a product manager for Apple and you want to know what people think about 'iPhones' - a quick scan with the right tools will tell you:</p> <ul> <li>How many people or sites discussed iPhones/smartphones.</li> <li>How the volume of iPhone/smartphone discussions trended over time.</li> <li>Where the iPhone/smartphone discussions happen.</li> <li>How people feel when they discuss iPhones/smartphones.</li> <li>What they are saying - over all time and at specific intervals - in the context of iPhone/smartphone discussions.</li> </ul> <p>As a market analyst, the information is useful for correlating with sales, stocks, etc. - and for understanding macro trends in order to anticipate the near future.</p> <p>As a product portfolio manager, the information is useful for understanding customer satisfaction.</p> <h3>Care for your audience: Customer service with social listening</h3> <p>In fact, customer service is one of the major applications of social listening technology.</p> <p>Tools like Sprinklr and SparkCentral are designed to power customer service for enterprise organizations.</p> <p>By observing real-time lists of unhappy customer data, organizations can react quickly to address customer concerns, as well as observe major themes that arise from dissatisfied customers.BIS</p> <p>For example, remember #Bendgate? That crazy thing where iPhones could bend? Yeah, that all started in September 2014. See this chart displaying over 450,000 mentions between September 20 and 30. </p> <p>The mentions are either of #bendgate (green) and “iPhone NEAR/5 bend” (purple; NEAR/5 is a Boolean search operator that will only allow results where the second word is within 5 words of the first):</p> <p><strong><img src="https://assets.econsultancy.com/images/0006/8567/social_listening_chart.png" alt="" width="975" height="396"></strong></p> <p><em>Courtesy of <a href="http://infegy.com/">Infegy Atlas</a> - our preferred social listening tool.</em></p> <p>Notice that things don’t really pick up until about the 25th, but the mentions begin streaming in on the 23rd.</p> <p>Before the hashtag really caught fire, mentions of iPhones bending were higher. In fact, there are mentions of iPhones bending as far back as the 21st (though it’s not visible on this chart because of low volumes).</p> <p>Savvy use of a social listening tool would allow those who manage customer satisfaction to watch this theme of bending iPhones unfold before it went viral.</p> <p>You can beat the press to the punch with social listening.</p> <h3>Activate your audience: Targeted advertising and influencer activation with social listening</h3> <p>Top-tier social listening tools should provide you with unique IDs of social users, where possible.</p> <p>Typically, this is only feasible on Twitter and Instagram, and sometimes other less-used social platforms like Reddit and Weibo.</p> <p>For example, see this list of targeted Twitter users developed using <a href="https://manageflitter.com/">ManageFlitter</a>: <a href="https://twitter.com/evanpdunn/lists/boston-cmos">CMOs in Boston</a>.</p> <p><img src="https://assets.econsultancy.com/images/0006/8562/Screen_Shot_2015-11-02_at_13.52.35.png" alt="" width="565" height="303"></p> <p>Once you have unique IDs of targeted users - especially ones who have strong affinity for your brand - you can do two things:</p> <p><strong>1. Deliver advertising to fans via Custom Audiences</strong></p> <p>Twitter and Facebook both offer advertising solutions that involve bulk uploading unique IDs of users in order to deliver advertising to only them. These are called <a href="https://econsultancy.com/blog/64980-put-your-email-list-to-work-facebook-custom-audiences/">Custom Audiences</a>.</p> <p>Say you develop advertising attribution software and are trying to sell to marketing execs.</p> <p>You are based in Boston - or are boosting sales efforts there - so you want to develop awareness by placing your message in front of marketing execs in that area.</p> <p>With that “CMOs in Boston” list I generated through ManageFlitter, you could deliver advertising to these users. If your list isn’t large enough (over 500 users), then you could broaden your search to include other qualifying terms.</p> <p>If you have users signifying positive sentiment - or are using more sophisticated NLP software that looks at purchase intent - you could pull the user results from Twitter and advertise upsell products to them.</p> <p>DISCLAIMER: According to Twitter's Terms, exporting lists from searches such as the one described above in order to upload them to Custom Audience for advertising targeting is not viewed favorably. Legal ramifications may result.</p> <p>You can read more about it in the comments below.</p> <p><strong>2. Activate influencers</strong></p> <p>Most social listening tools have some level of focus on telling you who your “influencers” are.</p> <p>These are the power users - the ones who advocate strongly for your brand and who have a good degree of visibility online.</p> <p>Much like Custom Audiences for advertising, you can act on the user lists generated by these tools to encourage influencers to talk more about your brand.</p> <p>Some brands offer these influencers special promotions in exchange for posts; others invite them to exclusive events.</p> <h3>Benchmark against your competition: Landscape analysis with social listening</h3> <p>Share-of-voice is a common measure of competitive standing among major brands. With the right tools, you can develop macro SoV insights, as well as SoV by geographic segment, channel segment (News, Forums, Blogs, etc.) and demographic segment (gender, age, etc.).</p> <p>You can also develop sentiment benchmarks. Perhaps your brand is getting largely negative feedback, so you develop campaigns designed to encourage positive sentiment online.</p> <p>You can set incremental measures of success so that, in six months or a year, you’re on top of the positivity game.</p> <p>Marketers can also use social listening to track negative press about competitors.</p> <p>In one instance with a higher education client, we knew about a disastrous press situation for our client’s competition before our client did. Many social listening tools offer alert functions that make this possible.</p> <h3>Optimize your campaigns: Real-time monitoring with social listening</h3> <p>Content marketing for enterprise is often focused around key phrases, tag lines or hashtags.</p> <p>Social listening tools allow you to track the spread of these beyond your own paid/earned/owned efforts to promote them.</p> <p>You can also measure the saturation of your hashtags and campaign terms among the overall buzz about your brand.</p> <p>For example, for a major telecom client we were able to see the points at which their online conversation was dominated by a new hashtag they had launched.</p> <p>This allows them to optimize their content based on the apparent efficacy of these times.</p> <p>And finally, you can deploy social listening tools to measure the efficacy of your campaigns not only in volume, but also in sentiment.</p> <p>After all, what good is buzz if it’s bad?</p> <p><em>Where to learn more about social listening tools:</em></p> <ul> <li><a href="https://econsultancy.com/blog/65958-social-listening-in-2015-top-five-opportunities/"><em>Social listening in 2015: top five opportunities</em></a></li> <li><a href="https://econsultancy.com/blog/64864-16-essential-social-media-management-tools-we-love/"><em>16 essential social media management tools we love</em></a></li> </ul> tag:www.econsultancy.com,2008:BlogPost/66948 2015-09-21T13:50:00+01:00 2015-09-21T13:50:00+01:00 What a Dislike button on Facebook could mean for brands Patricio Robles <p>In a Q&amp;A session last week, Facebook founder Mark Zuckerberg revealed that his company is working on a way for Facebook users to express different emotions.</p> <blockquote> <p>People have asked about the 'dislike' button for many years. Probably hundreds of people have asked about this, and today is a special day because today is the day that I actually get to say we are working on it, and are very close to shipping a test of it.</p> </blockquote> <p>Zuckerberg didn't reveal much more than that. He did indicate that the intent of this new button would not be to facilitate "down vote[s]." Instead Facebook would be providing a way to avoid liking a piece of content when it might be seen as insensitive. For example, liking a Facebook post about someone's passing is considered by many to be an awkward act.</p> <h3>A Dislike button could change Facebook</h3> <p>A Facebook in which users can do more than Like content could create a very different environment for brands using the social network.</p> <p>After all, even if Facebook intends for its yet-to-be-named Like alternative to be used as a mechanism for disliking content, the addition of a new way for users to express their emotions about content will likely require brands to rethink some of the ways in which they interact with Facebook users and how they evaluate their efforts on Facebook.</p> <p><img src="https://assets.econsultancy.com/images/resized/0006/7194/8155144549_f82feb1d2c_b-blog-flyer.jpg" alt="" width="470" height="157"></p> <h3>Content creation</h3> <p>A dislike button of some kind could complicate content marketing campaigns on Facebook, as brands would need to consider the possibility that users on the social network who don't like their content could use the new button to express their feelings.</p> <p>Right now, users can do this by posting a comment, but that takes effort and there's no visible figure on Facebook that displays the number of comments deemed to be negative.</p> <h3>Metrics</h3> <p>Beyond the possibility that the new button will come to be used as a down voting mechanism in practice – a true Dislike button whatever it's called – brands might need to adjust how they measure the conversations around their content.</p> <p>Instead of counting Likes, brands would need to count non-Likes and establish methods for evaluating the efficacy of their Facebook marketing campaigns in light of this new figure.</p> <p>For instance, brands might face the possibility of seeking to understand a piece of content that generates a large number of Likes and non-Likes.</p> <h3>A moot issue?</h3> <p>It's likely that Facebook is aware of the potential issues associated with a Like alternative and it's entirely possible it will keep this from complicating matters for brands.</p> <p>For example, it could decide to keep the Like alternative away from brand Pages. Or it could allow brands to decide whether or not the content they post would have anything other than a Like button.</p> <p>Whatever happens, Mark Zuckerberg's announcement that Facebook is looking to move beyond the Like is yet another reminder that even the largest social platforms are still subject to change significantly as they evolve.</p> <p><em>If you want to learn more about social from some of the most successful brands on the planet, get a ticket for the <a href="http://ecly.co/1EmHi7L">Festival of Marketing 2015</a> and check out the Social stage. </em></p> tag:www.econsultancy.com,2008:BlogPost/66789 2015-08-31T10:02:00+01:00 2015-08-31T10:02:00+01:00 We analysed 82 Econsultancy email subject lines and here’s what we learned Parry Malm <h3>The first challenge is to understand <em>why</em> some subject lines are good, and <em>why</em> some are bad.</h3> <p><a href="https://econsultancy.com/blog/66328-211-awesome-phrases-for-email-subject-lines-that-sell">Subject lines are important</a>. This isn’t news. But how can you quantify the 'goodness' and 'badness' of a subject line?</p> <p>Consider these two subject lines Econsultancy sent out earlier in 2015:</p> <ol> <li>Marketers, vent your frustrations here!</li> <li>Marketing Pain Points – share your stresses, get free report</li> </ol> <p>These two subject lines invited people to fill out a survey about <a href="https://econsultancy.com/blog/66623-three-remedies-for-your-marketing-pain-points/">marketing pain points</a>.</p> <p>Before we look at results, here's some similarities and differences between the two subject lines:</p> <ol> <li>Version A is shorter than Version B.</li> <li>Version A uses an exclamation point, while Version B doesn’t have any ending punctuation.</li> <li>Version A begins with a collective noun, while Version B begins with a proper noun.</li> <li>Version A uses the word ‘frustrations’, while Version B uses ‘stresses’.</li> <li>Version A has a comma to separate fragments, while Version B uses a hyphen.</li> <li>Version B offers a tangible incentive, while Version A doesn’t.</li> <li>Version B uses the word ‘free’, while Version A doesn’t.</li> <li>Version B uses jargon (‘pain points’), while Version A is more colloquial.</li> <li>Both use the imperative voice.</li> <li>Both use second person pronouns.</li> <li>….. and so on (I could go on – there are dozens of differences and similarities between the subject lines!)</li> </ol> <h3>Which subject line won – Version A or B?</h3> <h4>Version B won with a 7% open rate lift (at 99% confidence). Whoo hoo!<strong><br></strong> </h4> <p><strong>So now the important question: why did it win, and what can Econsultancy learn from this?</strong></p> <p>As mentioned above, there’s a litany of ways that Versions A and B differ. Dozens, or maybe even hundreds.</p> <p>Here are a few plausible hypotheses that this split test could answer:</p> <ol> <li>Longer subject lines are better than shorter.</li> <li>Using the word ‘free’ is good.</li> <li>Exclamation points are bad.</li> </ol> <p>And so on. The list could go on ad infinitum.</p> <p>This is just two subject lines! What happens when you look at dozens, hundreds or thousands of them?</p> <p>How can you tell which are the best… <strong>and understand why they’re the best so you can replicate the success?</strong> </p> <h3>What about these 10 subject lines?</h3> <p>These subject lines are the ones that experienced the biggest variation from the mean.</p> <table style="width: 517px;"> <tbody> <tr> <td width="379"><strong>Subject Line</strong></td> <td width="139"><strong>StDevs from Mean</strong></td> </tr> <tr> <td width="379">Upcoming training course I thought you'd be interested in</td> <td width="139">5.65</td> </tr> <tr> <td width="379">Where will marketers prioritise spending in 2015?</td> <td width="139">-2.59</td> </tr> <tr> <td width="379">eCRM and analytics - join a three-day masterclass</td> <td width="139">1.79</td> </tr> <tr> <td width="379">Training: content and SEO working together</td> <td width="139">-1.45</td> </tr> <tr> <td width="379">*drum roll* And the top 100 UK digital agencies are....</td> <td width="139">1.34</td> </tr> <tr> <td width="379">Adopting a lean methodology: training in agile marketing</td> <td width="139">1.22</td> </tr> <tr> <td width="379">Understanding the Customer Journey - get your free report</td> <td width="139">-1.20</td> </tr> <tr> <td width="379">What are marketers’ priorities in the year ahead?</td> <td width="139">-1.12</td> </tr> <tr> <td width="379">Understanding the Customer Journey - new report</td> <td width="139">0.91</td> </tr> <tr> <td width="379">Intensive training in developing a positive customer experience</td> <td width="139">-0.78</td> </tr> </tbody> </table> <p>(<em>Note: we normalised the results to determine the distance from the mean. This is to quantify variance that exists while controlling for other factors. A positive number means it was good, and a negative number means it was bad. The larger the number, the greater magnitude of its goodness or badness.)</em></p> <p><strong>Remember the differences we found between Versions A and B above?</strong> Let’s say we can find 10 differences between any two subject lines in the set of 10 above.</p> <p>This means that A is different from B in 10 ways, from C in 10 ways, and D in 10 ways, plus B is different from C in 10 ways, D in 10 ways... and so on.</p> <p>Suddenly you have a <a href="https://en.wikipedia.org/wiki/Combinatorial_explosion">combinatorial explosion</a> of millions of hypotheses, all of which could be correct. Discovering which is ‘the one’ is like finding the proverbial needle in the haystack.</p> <p><strong>So, what’s an email marketer to do?</strong> </p> <h3>Here’s how you can quantify <em>why </em>some subject lines work and others don’t</h3> <p>You read a subject line as a chunk of text and emotionally respond to it. That’s how language works. You ingest an entire sentence, have a subconscious response, and act upon your impulse. <strong>Email subject lines work when they pique particular emotions in recipients. </strong></p> <p><strong>But how do you quantify the emotions that resonate with your audience? </strong></p> <p>There are a couple of challenges here. First of all, hand-quantifying emotion is wrought with <a href="https://en.wikipedia.org/wiki/List_of_cognitive_biases">cognitive biases</a>. And second, it’s a time-consuming and cumbersome task.</p> <p>We ran Econsultancy’s data through our email subject line sentiment analysis tool - <a href="https://phrasee.co/introducing-phrasee-pheelings/">Phrasee Pheelings</a>. It’s the world’s first technology that measures the emotional power of any of your previous subject lines.</p> <h3>Here’s the bad news for Econsultancy</h3> <p>Econsultancy tends towards two distinct emotional clusters in their subject lines, loosely described as:</p> <ol> <li>Direct and devoid of personality (i.e. ‘<em>The Consumer Conversation - new report’</em>) - in blue below</li> <li>Mid-level familiar and curious (i.e. ‘<em>How healthy is your content strategy?’</em>) - in pink below</li> </ol> <p><img src="http://assets.econsultancy.com/public/imgur/W5Cflzf.jpg" alt="Phrasee Pheelings 2 cluster" width="600" height="600"></p> <p>The response variance of the two clusters was noticeable, though small. Why is this?</p> <p>When marketers have to write loads of subject lines in their day jobs, they either</p> <ol> <li>follow the status quo; or</li> <li>follow gut instinct.</li> </ol> <p>Both of these strategies run many risks, not least of which are:</p> <ul> <li>If the status quo subject lines weren't good in the first place, you’re just polishing a turd. You aren't progressively optimising your language decisions.</li> <li>If unquantifiable assumptions (like the Versions A and B comparison above) drive your gut instinct, you are making uninformed decisions.</li> </ul> <p>People tend to do the same things over and over, because “if it ain’t broke, don’t fix it.” The results don’t get worse, but also they don’t sustainably improve.</p> <p><strong>This is a missed opportunity.</strong></p> <h3>Here’s the good news for Econsultancy</h3> <p>Econsultancy’s two main emotional clusters don’t vary much, and the results don’t either. But this is an opportunity in itself!</p> <p>There are limitless possible emotional combinations of language available. <strong>Knowing which work best for your audience is impossible without testing. </strong></p> <p>The key is to get <a href="https://econsultancy.com/blog/65885-how-304-email-marketers-split-test-their-subject-lines-spoiler-alert-it-s-usually-badly/">split testing.</a> But don't test pointless things like changing a single word. That’s a tiny learning point that won’t tell you much about your audience. It will just make you buy a thesaurus.</p> <p><strong>Instead, test out big things. Like the overall emotional power of your subject lines. </strong></p> <p>And do it in a methodological, robust, quantifiable manner.</p> <h3>So, Econsultancy, here’s our advice to you:</h3> <p>When writing your subject lines, test the emotional impact of your subject lines.</p> <p>You’ve tested out two emotional clusters ad infinitum, and the results aren’t bad. But what about the clusters you haven’t tried yet?</p> <p>For example:</p> <ul> <li>What about high urgency/’FOMO’ sentiments? There’s few like this in the sample set... but what if they're key drivers for your audience? In orange below.</li> <li>How about more familiar tones? Econsultancy has a friendly brand voice, yet that doesn’t translate in your subject lines. In purple below.</li> <li>And, what about high levels of curiosity and ambiguity? In green below.</li> <li>And so on, and so on... there's limitless combinations to consider.</li> </ul> <p><img src="http://assets.econsultancy.com/public/imgur/jbWo5Ij.jpg" alt="Phrasee Pheelings Untested clusters" width="600" height="600"></p> <p>To complicate matters – none of these emotions exist in isolation. You can have something that’s both urgent and familiar, for example. Or something that’s not that urgent, but has middling values of both familiarity and curiosity.</p> <p>The challenges are:</p> <ul> <li>to quantify what emotions to test;</li> <li>to generate subject lines that adhere to these emotional clusters; and</li> <li>to measure on an ongoing basis response, while controlling for external factors.</li> </ul> <p><strong>When testing out emotional clusters, some will work, and some won’t. </strong></p> <p>Your audience responds to the emotional power of your subject lines. Until you figure out what combination of emotions work the best, you’re using suboptimal subject lines.</p> <p><strong>The take-home point is this: </strong></p> <p>Small language sets are suboptimal if you don't quantify their success or failure. You don’t know if what you’re doing is good… or if you’re missing an opportunity.</p> <h3>Big thanks go to…</h3> <p>What I like about the good people at Econsultancy is they’re always up for <a href="https://econsultancy.com/blog/65263-3d-printing-check-me-out/">trying new things</a>… and sharing what they learn with their audience. They love innovation and that’s why they’re such a well-respected group.</p> <p>So thanks a lot to the Econsultancy team for sharing their data with us here at Phrasee, and for being good sports through the process. </p> <h3>And it’s not just us who think this is important…</h3> <p>When writing this blog post I received the following email from Econsultancy’s <a href="http://www.festivalofmarketing.com/index.cfm">Festival of Marketing.</a> Check out the subject line:</p> <p><img src="http://assets.econsultancy.com/public/imgur/L2jkDhL.png" alt="" width="600" height="420"></p> <h2> </h2> <h3>Emotional power is what makes marketing awesome.</h3> <p>We all know that emotion drives response. That’s not the opportunity.</p> <p>The opportunity is to quantify emotion… then optimise based on the results… and then <strong>profit from better subject lines.</strong></p> tag:www.econsultancy.com,2008:BlogPost/66632 2015-06-29T14:48:00+01:00 2015-06-29T14:48:00+01:00 Five creative uses of data from Cannes Lions Celina Burnett <p>As brands and agencies we have access to increasing volumes of data and we are using these to make our customer experiences more relevant, more interactive and more effective.</p> <p>The inaugural Cannes Lions Innovation festival which took place this week reflects this shift and shows us how brands and agencies across the industry are achieving this and provides a glimpse of innovations to come. </p> <p>More than 600 entries were submitted to the new festival under the Creative Data Lion award, which will be a permanent fixture at the annual Cannes Lions International Festival of Creativity to celebrate the intersection of data, technology and ideas.</p> <p>The award covers a broad spectrum of categories, including the use of real-time data, visualisation and technology to enhance creatives and consumer engagement, but a key trend from this first year of submissions has been the use of data and insight to drive targeting and personalised experiences, in particular the use of real-time and social media data.</p> <p>There was no Grand Prix awarded for the Creative Data Lion this year. No entry stood head and shoulders above the rest, however individual category winners set a benchmark for industry and for next year’s entrants.</p> <p>These include: </p> <h3>Netflix: Inspired By You (France)</h3> <p>Netflix launched in France in September 2014. To ensure success the business used text mining and machine learning to analyse two years’ worth of digital conversations.</p> <p>This analysis provided granular insight into the French population’s needs, desires and interests to inspire relevant and contextual creatives and messaging for Netflix’s outdoor campaign, a campaign which the organisation overlaid with real-time contextual data such as weather to further tailor messaging and increase relevance.</p> <p>The campaign reached 120m and increased brand awareness from 25% to 68% in just three months.</p> <p><iframe src="https://www.youtube.com/embed/W76Sc6lmOvA?wmode=transparent" width="425" height="350"></iframe></p> <h3>Tennessee Department of Tourist Development: One Minute Vacation (USA)</h3> <p>To encourage local tourism, Tennessee Tourism used demographic and travel data to identify 11 local markets that were within a short drive of Tennessee’s biggest attractions.</p> <p>They mapped out 24 specific road trips from these markets with minute-long video banner ads to illustrate and market each. Using social media location, interest and conversation data these ads were delivered as part of a highly targeted campaign.</p> <p>When the user clicked-through they were shown the exact roads to take to get from their location to the destinations advertised.</p> <p>In five weeks the campaign generated 17,000 hours of video views and resulted in 26,000 holiday guide downloads, an increase of 27% over the previous year.</p> <p>The campaign contributed to a 10.7% increase in travel volume to Tennessee that year.</p> <h3>EA Sports: Madden Giferator (USA)</h3> <p>The makers of Madden, the NFL-based video game, had seen its popularity decline in recent years.</p> <p>The business was seeking to reverse this by tapping into the excitement and team rivalry of the NFL season to reach the 157m NFL fans in a real-time and relevant way with the creation of the Giferator.</p> <p>This dynamically generated GIFs in real-time, layering headlines and stats from NFL games as they happened with video game artwork featuring the relevant players.</p> <p>These GIFs were then distributed across sites where fans were known to be looking for game updates and were targeted according to team loyalties.</p> <p>Fans were also given the opportunity to create their own GIFs, resulting in more than 420,000 GIFs shared online by fans and players.</p> <p><iframe src="https://www.youtube.com/embed/spPjhKwthHI?wmode=transparent" width="425" height="350"></iframe></p> <h3>Mercedez Benz: Crossword Targeting (Japan)</h3> <p>Mercedes-Benz sees 8.5m visits to their website each month, but few are hot prospects.</p> <p>To reach their target audience Mercedes-Benz used signals from search data to deliver targeted display advertising.</p> <p>In total it identified 4.6m combinations of search keywords that indicated the user might be a hot prospect to deliver ads against. For example, search keywords which when combined indicate a car accident or repairs.</p> <p>The click-through rate for the campaign was 1070% higher compared to traditional targeting.</p> <h3>Melanoma Patients Australia - Melanoma Likes Me (Australia)</h3> <p>More than 1,500 Australians die from melanoma each year, it is the most lethal cancer for 15-30 year olds.</p> <p>The Melanoma Likes Me campaign was created to raise awareness of the dangers associated with sunbathing amongst this younger and harder to reach audience.</p> <p>A real-time tool was used to respond to popular hashtags such as #sunkissed, #tanned and #beachside and geo-located images on social media with likes and messages from user _melanoma.</p> <p>Information on how to check for melanomas was also provided when users clicked through.</p> <p><iframe src="https://www.youtube.com/embed/AGL3Pz_98xQ?wmode=transparent" width="425" height="350"></iframe></p> <p><em><strong>Make sure to check out <a href="http://ecly.co/1L5KkCk">The Masters of Marketing</a>, Econsultancy’s new awards hosted in partnership with Marketing Week.</strong></em></p> <p><em><strong>Covering the key marketing disciplines and industry sectors, don’t miss out on this opportunity to highlight your work.</strong></em></p> tag:www.econsultancy.com,2008:BlogPost/66571 2015-06-12T17:06:00+01:00 2015-06-12T17:06:00+01:00 Marketing analytics: what your web traffic says about your business Lori Goldberg <h3>Separate perception from reality</h3> <p>Often, a business’ perception of its customer is very different than the customer’s actual behavior. Fortunately, analytics can support or challenge your perceptions and help you build a profile of the customers that you attract. <br> </p> <p>For example, you can determine: </p> <ul> <li> <strong>Geographic location</strong> – Pinpoint city, regions, states or country</li> <li> <strong>Demographics and interests</strong> – Combine remarketing data with your site analytics to reveal age, gender, and special interests by category</li> <li> <strong>Time of day</strong>: Do your users peak at certain times of day? What’s triggering this behavior? </li> </ul> <h3>Know if your traffic is real</h3> <p>Perhaps you’re seeing spikes of spam bot traffic hitting your site and generating false positives about your foreign audience.</p> <p>High levels of traffic from Netherlands, Ghana, China, Russia and other faraway lands coupled with high bounce rates are likely an indication of bots hitting your site or gaming your analytics code to appear as a visitor, even if they’ve never been to your site.</p> <p>The most common tactic occurring now is referral traffic sent to you from suspicious sites. Check your referral reports for names such as “Simple-Share-Buttons,” “Event-Tracking.com” and “Get-Free-Traffic.”</p> <p><img src="https://assets.econsultancy.com/images/0006/4026/referrer-spam.png" alt="" width="605" height="268"></p> <p>A link to these sites will appear in your reports - but don’t click on them – they may be a gateway to a computer virus, Trojan, or a black hat SEO firm advertising their services.</p> <p>You can eliminate this data from your reports by creating filters, though it requires a diligent effort to stay one step ahead of the bad guys as the names often change.</p> <p>In the end, you may no longer have the international following that you felt you had, but <strong>your data will be clean and tell a more accurate story of your business.</strong></p> <h3>Determine the strength of your brand</h3> <p>“Direct Traffic” is any traffic that comes to you via a direct link or URL and not via more choice-driven responses such as a search engines, social media, or advertising.</p> <p>A healthy percentage of direct traffic will indicate that consumers are familiar with your brand and likely type it directly into their browser. Look for this percentage to grow over time as your brand benefits from advertising, repeat business, and consumer memory.</p> <p>Additionally, your brand name should be a top 5 keyword for your business as this indicates strong brand memory in the minds of consumers searching for you. If your brand name is not strong enough to break your top 5, consider a branding campaign or public relations to provide lift and grow name recognition.five</p> <h3>Long tail keyword data</h3> <p>The keyword choices that consumers use to find your business can tell you a lot. They may misspell your company name, associate you with a competitor, know you only by a product that you sell, or search for an employee or spokesperson by name.</p> <p>By collecting this long tail keyword data, you can better understand how your customers know you, what they call you, what they want from you, and perhaps where they’re confused about your brand.</p> <p>It’s important to note that in Google Analytics, some keyword information may be categorized as “not set” meaning the source is not available. It can be frustrating to marketing analytic gurus to not have this window into their data.</p> <p>“Not set” typically means the user is searching privately with a cloaked browser or is logged in to a search engine and has chosen to not have their behaviors tracked. </p> <p>In AdWords, however, this data becomes available as a part of Google’s advertising services. If you encounter “Not Set” data, substitute with AdWords data when available.  A few years ago, Google set default browser search settings to private; however, Yahoo/Bing searches provide keyword-level data within <a href="https://econsultancy.com/blog/65647-a-thoroughly-entertaining-beginner-s-guide-to-data-and-analytics">Google Analytics</a>.</p> <p>Also, be aware of fake organic keywords that appear in your reporting and make their way into your data. Ghost referrals are keywords that have utilized a weakness in Google Analytics to appear in your reporting.</p> <p>These suspicious keywords may appear as “Get Free Traffic” or something similar in an attempt to play like an advertisement for the company that placed the term in your reporting. The terms are typically obvious and have no relevance to your business.</p> <p>Image credit: <a href="http://blog.raventools.com/stop-referrer-spam/">raventools</a></p> tag:www.econsultancy.com,2008:BlogPost/66494 2015-05-26T10:23:59+01:00 2015-05-26T10:23:59+01:00 Computers can rap better than Kanye. What does this mean for marketers? Parry Malm <blockquote> <p>My name is Parry and I’m here to say / Natural language processing is worth a play / There’s technology out there making rhymes / It makes you money all the time.</p> </blockquote> <p>OK, that was the worst rap ever. I won’t quit my day job, I wasn’t destined to be a rapper. Even though I grew up on the mean streets of Vancouver’s suburbs, the ancestral home of gangsta rap.</p> <p>Luckily, there is an emerging field within artificial intelligence that will prevent me every having to punish you with my sick rhymes ever again.</p> <h2>Which verse is by the famous rapper?</h2> <p>One of the lyrics below was written by a famous rapper. One was written by a computer. And one was written by me.</p> <blockquote> <p><strong>Version A:</strong><br>I gots big love for you, fa sho<br>I know, you know, we know<br>Truth be told I loves u for real yo<br>Our love blows up like volcanos</p> <p><strong>Version B:</strong><br>I know, you know, we know<br>Cause I slow ya roll and you know it bro<br>Let me really tell you how it should be told<br>You think not? You could see though</p> <p><strong>Version C:</strong><br>What we have is real <br>Wanna shout it from the mountains baby<br>That I love you, but if they don't hear me<br>I know, you know, we know</p> </blockquote> <p>Give those three verses a read, and… let’s take a poll. Which one is the real rapper?</p> <p><strong><a href="http://directpoll.com/v?XDVhEtJ2CxqI41AHyH4QxPaAS1nN1Tt" target="_blank">Click here to vote</a></strong></p> <p>Let’s not deconstruct the deep, philosophical meaning of these lyrics. We’ll leave that for the unemployed English graduates of the future.</p> <p><strong>Let’s look at the linguistic construction of the phrases:</strong></p> <ul> <li>Versions A, B and C all rhyme, or at least follow a rhyming pattern.</li> <li>Versions A, B and C all follow a clear rhythmic <a href="http://en.wikipedia.org/wiki/Pentameter" target="_blank">pentameter</a>. </li> <li>Versions A, B and C are all plausible verses in a rap song.</li> </ul> <p>But only one of them was written by a famous rapper.</p> <p>Now – to avoid spoilers and cheaters – here’s a brief musical interlude to make sure you don’t look at the answers before you fill out the poll above.</p> <p>This interlude is brought to you by an amazing rapper whose name (Childish Gambino) was <a href="http://genius.com/posts/125-Donald-glover-interview-what-the-fuck-is-a-childish-gambino" target="_blank">computer generated</a>, which seems fitting... </p> <p><iframe src="https://www.youtube.com/embed/ExVtrghW5Y4?wmode=transparent" width="560" height="315"></iframe></p> <h2>Here’s my prediction of the results:</h2> <p>Bear in mind the results are tallied in real time and I’ve no idea what the results will be at the time of writing this blog.</p> <p>But, here's what I predict the results will say: <strong>none of the 3 verses will be a clear winner.</strong></p> <p>I reckon all three will get some votes, unless this vote gets tainted by rap nerds who look the lyrics up. Assuming that won’t happen, there will be no clear winner. So don’t be a rap nerd, let’s keep this an untainted poll based upon your own gut response to the lyrics.</p> <h2>OK, so, which one was written by a real rapper?</h2> <p><br><a href="http://directpoll.com/r?XDVhEtJ2CxqI41AHyH4QxPaAS1nJnEuYEnGdzeDuB" target="_blank"><strong>Click here to see the poll results in real time</strong></a></p><p><strong>Version A</strong> was created by a straight gangsta homeboy who is a devilishly handsome, self-described digital anarchist. Yep, that verse was my own amazing attempt at being a rapper. I’m the next Macklemore, that’s clear.</p> <p><strong>Version B</strong> was created by an algorithm devised by Finnish computer scientists.</p> <p><strong>Version C</strong> was created by Talib Kweli in his song, “<a href="https://www.youtube.com/watch?v=RTbfx4N8lLE" target="_blank">We know</a>.”  </p> <h2>How did the Finnish scientists generate their verse?</h2> <p>They used “<a href="http://en.wikipedia.org/wiki/Natural_language_processing" target="_blank">Natural Language Processing</a>,” a field of <strong>artificial intelligence that has incredibly exciting commercial potential.</strong></p> <p>To quote their paper:</p> <blockquote> <p>Our work lies in the intersection of the areas of computational creativity and information retrieval. In our approach we assume that users have a certain concept in their mind, formulated as a rap line, and their information need is to find another line, or a set of lines composing a song.</p> </blockquote> <p>In other words, they take existing rap lyrics. They interpret themes within the text. Then, they determine the number of syllables, and the rhyming and pentameter constructs. Then they computationally select relevant lines to make an effective verse.</p> <p>Here’s an example of another lyric they created, alongside the source artist:</p> <table style="width: 400px;" border="0"> <tbody> <tr> <td><strong>Line</strong></td> <td><strong>Original Artist</strong></td> </tr> <tr> <td>For a chance at romance I would love to enhance</td> <td>Big Daddy Kane - The Day You're Mine</td> </tr> <tr> <td>But everything I love has turned to a tedious task</td> <td>Jedi Mind Tricks - Black Winter Day</td> </tr> <tr> <td>One day we gonna have to leave our love in the past</td> <td>Lil Wayne - Marvin's Room</td> </tr> <tr> <td>I love my fans but no one ever puts a grasp</td> <td>Eminem - Say Goodbye Hollywood</td> </tr> </tbody> </table> <p>Note how each lyric is from a different artist. Yet, they all share the same theme (love), the same “flow,” or pentameter, and they all rhyme. And they don’t just rhyme at the end of each phrase like a junior-high love letter - the entire stanza has a human-sounding rhythmic pentameter.</p> <p>The scientists measured their success on a metric they call “rhyme density.” This is a statistic that measures how many rhymes exist both within a stanza, but also within a given line, and is a metric for overall flow of a rap verse.</p> <p><strong>Their algorithm out-performs the rhyme  density of regular rappers by 21%.</strong> <a href="https://www.youtube.com/watch?v=Fr0A7TofowE" target="_blank">Sheeeeeit</a>.</p> <p><img src="http://assets.econsultancy.com/public/imgur/a6VXXJD.jpg" alt="" width="520" height="347"></p> <p>By the way, the rapper in the research with the highest rhyme density was Inspectah Deck, one of the original Wu’s who was never very commercially successful on his own. Which is a shame, as he is a fantastically talented rapper. Listen to this song and take note of his rhyme density. <strong>The algorithm even out-performed him.</strong></p> <p><iframe src="https://www.youtube.com/embed/W-9e3hCvE28?wmode=transparent" width="560" height="315"></iframe></p> <p><br>The full research paper can be found <a href="http://arxiv.org/pdf/1505.04771v1.pdf" target="_blank">here</a> (PDF.)</p> <h2> <br>That’s “dope,” as the kids say, but what does it mean for marketers?</h2> <p>Human language is an incredibly complex problem to solve computationally. I sincerely doubt a computer will be able to write better lyrics than Big L (RIP), for example, notwithstanding his helpful translation of Ebonics.</p> <p>But within small language spaces, for example AdWords, email subject lines, or calls-to-action on websites, there is HUGE potential for the artificially intelligent optimisation of language.</p> <p><iframe src="https://www.youtube.com/embed/TMeFcVHNT1Q?wmode=transparent" width="560" height="315"></iframe></p> <h2>Here’s a real marketing use case: email subject line optimisation </h2> <p>Email subject lines are, in essence, a set of words, phrases and sentiments that have a specific, measurable outcome – that is, making someone open and act upon your email.</p> <p>For any given subject line, there are <strong>trillions</strong> (for real, yo!) of possible ways to convey your message. The human brain, however, is creatively limited… and is fundamentally not objective. Humans can’t think of even a minor fraction of all the possible variants.</p> <p>Using artificial intelligence you can consider huge volumes of data – exponentially more than a human brain can – and then apply <strong>artificial intelligence to optimise marketing language generation</strong>.</p> <p>There is software out there that does this right now. Full disclosure: my company Phrasee does this, so I'm obviously biased, but there are other companies who are commercialising this emerging technology as well. It’s a nascent but fast growing market. Early adopters of this technological innovation are reaping the benefits.</p> <h2>The future of copywriting is artificial</h2> <p>Many marketers are precious about their words being used, like they’re the <a href="https://www.youtube.com/watch?v=41qC3w3UUkU" target="_blank">next 2Pac of the marketing world</a> (NSFW)... and are unwilling to believe that artificial intelligence can out-perform them.</p> <p>Perhaps it’s job security worries. Or perhaps it’s because they don’t understand the power of natural language processing technology.  Or perhaps it’s just good old inertia and risk-aversion.</p> <p><strong>The main point is this:</strong></p> <p>If you’re precious about your own words being used in your marketing, regardless of the quantitative results, then carry on as you are. No worries, <a href="https://www.youtube.com/watch?v=6FwYqAg8wUE" target="_blank">this blog ain’t about you</a>.</p> <p>If you would rather use artificial intelligence to improve your marketing language and increase your response rates, then you could be the next <a href="https://www.youtube.com/watch?v=0LNoZraRTqg" target="_blank">straight marketing gangsta</a>.</p> <p>If you take away one thing from this blog post, take away these wise words, which were (obviously) written about natural language processing being a nascent technology with incredible potential.</p> <blockquote> <p>Uh 1-2-3 that's how it be<br>Somebody gotta be doubtin' it<br>It was me Michael T.T. handing out freebies<br>Let's party everybody bounce with me</p> </blockquote> <p>BTW: this was a computational combination of Outkast, KRS One, Ghostface Killah and 50 Cent. Just in case you were curious. I’ve no idea who Michael T.T. is, but I'm sure he's a nice guy.</p>