Where do you start?

Describing how his team overcame these issues, Robertson said that one of the main considerations was identifying who the end users were and how they needed to use the data.

It also required analysis of:

  • The devices on which people would access the data.
  • The complexity of data needed by the end user.
  • How data can be delivered in real time.
  • How to create systems that can be used across the organisation.

Therefore, Robertson’s team had to build personas around the types of people who would be using the data and their individual requirements.

Here’s a run through of two of the personas Action for Children came up with…

Executive management team

Requirements:

  • The executive team has overall responsibility for the entire organisation, which also involves an element of fundraising.
  • To stay on top of things the team need to know what’s happening at all times within the organisation.
  • Decisions affect thousands of staff members, but need to be made quickly.
  • Generally on the move, so need to be able to access data systems from mobile devices.

Solution:

  • The data team created a self-service analytics visualisation tool. This is a key element as the executive team doesn’t have time to scrape through spreadsheets or databases.
  • It is a cloud-based system connected to a big data engine.
  • Security is a vital feature so there are rules in place for data access.
  • A more intensive, in-depth system is available in the office, and the executives can also request detailed reports and analysis.

Director of fundraising

Requirements:

  • In charge of all fundraising at Action for Children, which includes nine separate teams.
  • The director needs to see the big picture, but also understand the deeper granularity.
  • Access to a predictive analytics system is vital, as it helps with creating new campaigns.
  • Requires access to real time analytics, however this is made more difficult as the director of fundraising is responsible for the organisation’s largest database.
  • Her team generally isn’t IT focused, so data needs to be available in simple formats.
  • This function relies heavily on fundraisers inputting data correctly, so database cleansing is complex and costly.
  • Fraudsters often target charities, so the director requires a system of triggers to identify and raise alerts when this occurs.

Solution:

  • The data team designed a robust system of analytics and triggers that was available across different devices.
  • It enables the collection and storage of IP addresses that have been flagged during payment processes, in real time.
  • Allows real time checking against known fraudulent IP addresses. This includes a big data engine used to analyse historic gift aid declarations.
  • The system offers theme-based analysis for supporter selection. This helps to understanding the propensity of people to donate depending on the theme of a campaign, which is important for optimising fundraising initiatives.
  • It also allows campaign analysis based on demographic data, so the fundraising director can forecast the likelihood of certain people to donate or sign up to direct debits.
  • When campaigns are live the director has access to real time online donation analysis coupled with traffic stats. The charity can’t waste time running a campaign that isn’t bringing in money, so this helps to flag up initiatives that aren’t performing.
  • There is also a new big data engine that helps to reduce cost of data cleansing, better segment analysis, and self-service interactive analytics.

Picking the solutions

Finally, Robertson ran through his top tips for unlocking the power of data analytics, including working with vendors:

  1. Investigate the market place. Check forums and websites to make sure you know the best solutions available for your needs.
  2. Test products with your own sample data. Don’t just use the vendors data, you need to make sure data platform are compatible with your own database.
  3. Plan for the future. Action for Children is planning for a five-year roll out, so the technology has to be bought with longevity in mind.
  4. Factor in how many users need access to new platforms. Look at how many people will be using new software platforms, as subscription models can impact investment.
  5. Roll new systems out team by team. Don’t go for it all at once as it magnifies teething issues.
  6. Excite people by using data visualisations. People can be turned off by data, so coming up with interesting visualisations will help to get buy in.
  7. Map the user and the type of data they need access to. This is essential in the planning stage.