systems of insight and engagement

Companies today are increasingly moving from a customer-centric vision to a customer-centric mandate – one overseen by a company’s most senior-level executives. Delivering on this promise and executing it at scale in a multi-channel world is difficult. Between your stores/branches, email system, loyalty program, call center, website, mobile app, DMP, social/adtech platforms and more – each containing valuable customer data – the challenge is daunting.

Customer data platforms (CDPs) hold the potential to help companies gain a deeper understanding of their customers by aggregating and synchronizing data sources into a single view that can, in turn, be used for conducting comprehensive customer analysis, creating and monitoring segments, and, potentially, taking action on the data.

I’ve spoken to many marketers about their future plans surrounding CDPs and customer data. I’ve found that many of them are so caught up in the project of bringing their customer data together and creating unified customer profiles, that they don’t consider how they will use that data until much later in the process. I think there is a lot of danger in this approach, as it often leads them to purchase two solutions (a “system of insight” and a separate “system of engagement”), when they should only be purchasing one combined solution.

In this blog post, I’ll describe systems of insight and systems of engagement and outline the pitfalls of keeping these two systems separate.

Systems of Insight and Systems of Engagement

Typically, your end goal when you purchase a CDP is to activate your CDP data (i.e., put it to use) to deliver more relevant experiences to your customers and prospects. Accomplishing that goal requires two types of functionality: insights and engagement. Often, these functions are performed by two systems.

  • System of insight: Aggregates customer/prospect data and enables analysis of that data.
  • System of engagement: Delivers experiences to customers/prospects in one or more channels and enables measurement of the outcome.

Organizations sometimes break up their CDP purchasing process into two parts. They may decide to purchase a CDP that serves as a system of insight initially, while they decide to purchase or connect it to system(s) of engagement six to nine months later after they have made progress with the first solution.

This is a short-sighted approach. These companies will ultimately be disappointed (and will likely have over-spent), as the separate systems won’t be able to work together to enable real-time activation, nor will they allow for activation/personalization at the 1-to-1 level.

Two Platforms = Two Brains

There are two main problems with separating CDP and personalization functionality into different systems.

1. Limitations of Segments

Systems of insight are primarily designed to evaluate and identify audience segments (e.g., high-value customers, those who have purchased leather boots in the past six months, etc.) and then pass the data to a separate system of engagement. Segment data is undoubtedly valuable in many situations (such as when sending an email to your high-value customers), but when the CDP can only pass segment-level data to another system for engagement, you’re not able to deliver 1-to-1 experiences or 1-to-1 content. As a result, your only option for greater relevance is a greater number of smaller segments mapped to specific campaigns, forcing you to trade off relevance with increased complexity.

2. Suboptimal Architecture

Having two platforms – a CDP and a separate personalization platform – effectively creates a technology environment with two “brains.” Systems of insight are a type of advanced, centralized brain full of customer data that generates insights from that data. The reality, though, is that to execute campaigns that engage audience members at a 1-to-1 level (and in real time), a system of engagement must also act as a brain. It needs to leverage machine learning to sift through the vast amount of customer data and make the best decision about which experience, content, promotion, message or recommendation to deliver to each individual at any given moment (e.g., a person who just landed on your website, is opening your most recent email, is calling into your call center, etc.) and in what channel.

It doesn’t make sense to have all of your customer data live in both of these systems, nor does it make sense to feed your system of engagement the limited data it receives from the system of insight in order to make a decision about which experience to deliver.

Common Use Case Example

To uniquely communicate to individual customers – at scale – requires a CDP that can both understand and activate at an individual level. Let’s look at an example through the lens of two different customers, Susan Jones and Susan Smith. Their profile data is seen below.

systems of insight and engagement

Let’s say you want to engage both of these consumers via an email campaign to remind them that they have loyalty points to spend. You might begin by putting both Susans into a segment (of customers that have outstanding loyalty points) and send them an email campaign with the same creative reminding them that they have loyalty points to redeem.

systems of insight and engagement

The campaign may have some impact, but it is not particularly relevant to each person. Instead, you might decide to make additional segments for each kind of category affinity. In this case, let’s say Susan Jones has a preference for seasonal furniture while Susan Smith has a preference for household goods. This would result in emails like these below where each recipient receives creative content associated to a known affinity.

systems of insight and engagement

These loyalty points reminder emails are more relevant and compelling because they have a more relevant banner and offer.

Trying to accomplish all of this with segmentation alone will quickly result in large, unmanageable number of segments you must set up in a CDP and then map those segments to a large number of campaigns in your ESP.

For instance, you would need to make a “have outstanding loyalty points” segment for every major product category (e.g., seasonal furniture + have outstanding loyalty points; household goods + have outstanding loyalty points, etc.) and map it to a matching campaign in your ESP. Likely, you won’t just have one campaign. You will also want to engage customers who are not in the loyalty program, who are active loyalty members, who are likely to churn, etc. More still, you may want to differentiate among customers not just by category interest and loyalty status, but also by their tendency to respond to offers (and what type of offer they respond to). With these complexities, the number of segments you must set up and manage grows and grows and quickly gets out of control.

And this is just one example on one channel, so you can see how the problem expands exponentially with different campaigns across channels. As a result, most companies end up curbing their ambitions and instead settle for some group-level relevance – falling way short of a customer-centric mandate.

Unlike CDPs that are solely systems of insight, Evergage can activate at the segment level and at the 1-to-1 level. Evergage can enable far greater individual relevance in every touchpoint and let a business scale with much greater simplicity.

With the machine-learning capabilities of Evergage’s real-time personalization and customer data platform, it is possible (and easy) to deliver truly personalized emails like this:

systems of insight and engagement

What I have outlined is just one use case example on one channel, but the same points apply across websites, mobile apps, call centers, in-store (kiosks/POS/clienteling), in-branch, and other channels.

Final Thoughts

Evergage is the only solution purpose built to be a unified CDP/personalization platform (a system of insight PLUS a system of engagement) that understands customers at an individual (and segment) level and then activates that data – at both levels – across your channels.

Why go to all the trouble of integrating and aggregating vast amounts of data at the individual level, if you’re not going to act on the data at the individual level?