Choosing the right data to get the best customer insights

The challenge facing CMOs since the dawn of marketing has been the need to understand their customers (both current and potential) and to set the direction of their business to respond to, engage with, and ultimately serve the requirements of their customers.

Early CMOs spent their time speaking with and listening to their customers and prospects through focus groups, exit interviews, as well as qualitative and quantitative market research.  In the last twenty years, these sources have been supplemented, then overtaken, and perhaps may one day be left behind, by newer sources of customer insight.

Today, CMOs are bombarded with more and more sources of information on the customer as data have become available from every consumer touch point.  This tidal wave of data led to technology companies focusing on the importance of “Big Data”, which wrongly gave the focus to the data itself as the end goal, rather than the requirements of the data and the insights it could provide.

The benefit of this heightened focus on data is that every single company is aware of the potential value of their data, and most are spending IT dollars on collecting and storing data on their customers.  For some, unfortunately, the actual value will not equate to the potential.  Many businesses are going to find that the hills do not have gold in them.  It should not be assumed that data monetization is a given.  A further problem is that data often gets stored in silos and there is a tendency for many companies to feel that the hard part is done once they have started storing the data.

In reality, there are two fundamental challenges that need to be overcome.  First, there needs to be much more focus on starting with the business question and result required from the data.  This allows the smart CMO, working with their Analytics/Insights team to determine how best to answer that question and, most importantly, which data sources can be used to best get to the answer.  Furthermore, it means you can then determine the incremental value from adding new data sources to answer the question.  Typically, the 80:20 rule applies, so that much of the value can be gained from using one data set, but using it well.  We view this as “Intelligent Data” instead of Big Data.

Aimia has worked with several of our grocery clients globally to optimize their digital flyer and this work is a great example of using intelligent data well.  Previously, an email was sent to customers each week telling them about the “biggest and best” promotions out of the thousands of deals in store that week.  Our grocery clients’ aim was to move to a personalisation communication which would tell people about the specific promotions which would potentially bring them into the store.  We realized that focusing on the SKU-level transactional data would be the best data set to answer this challenge.  So a customer who never buys soft drinks would not hear about the latest Coca Cola promotion in their message.  But they would hear about the top 6-10 promotions most likely to get them to head to this particular supermarket rather than the competition.  This 1:1 communication has driven 1%+ incremental sales through leveraging one dataset.

The second challenge that CMOs need to consider is how to integrate different datasets to provide a more advanced picture of the customer and lead to a more profitable end result.  The key is to really consider the value being produced from adding a different source of data.  In the grocery example above, one natural extension could be to layer in web browsing data to look at what items the customer may be interested in.  In the client example above, the return on investment for the additional work involved was not deemed to be strong enough to make this worthwhile.

Ignoring the merging of different company data assets, and focusing on the different data sources within a single company’s domain; the challenge becomes bringing together the disparate data into a unified view of the customer, or single customer view (SCV).  This process requires a significant cost beyond the technology, and companies which have invested in their Data Management and Analytics functions will be best placed to carry this out.  Often this will involve the breaking of silos rather than setting up new data feeds, and this will involve the typical business challenges of inertia, complexity and the need to have a sufficient business case.

On a case by case basis the value of joining two data assets within an organization may not stack up.  But, taking a higher level view, those companies and their CMOs who are able to recognise a customer across all interactions, not just transactions, will be the most successful through their ability to provide a much more targeted and personalized communication for each customer.

Prioritization of data sources

The principle of focusing on the lowest hanging fruit remains true.  This will differ by organization, but given the typical datasets a company will hold here is my starting point on different datasets and their relative importance.

1. Transactional
2. Communications (campaigns and responses)
3. Web browsing (on own website)
4. Mobile application (browsing, transacting)
5. Wider web data (including social)
6. Contact center
7. Location