Personalisation — What Works and What Doesn’t?

Last week Drapers, N Brown Group, East and Sailthru hosted a discussion on what really works and what doesn’t when it comes to personalisation. Hundreds of retail marketers and leaders joined the live webcast, where participants covered the hallmarks of a great online customer experience, what personalisation tactics customers truly value, the blockers keeping retailers from meeting their personalisation goals, and more.

You haven’t missed your chance on viewing the discussion — check it out here.

Many questions were asked from the audience — and we’ve taken a crack at giving our best advice and answers below. Feel free to let us know what you think on these topics by contacting us at

Q: Is investing in personalisation at the expense of optimising the core customer experience?

A: While multiple members of the webcast panel cited that they are working to get their core customer experience right and then dive deeper into personalisation, it is critical that retailers plan for CX improvements with personalisation in mind. When we think about personalisation, it’s always within the context of the greater customer experience and total business strategy, not simply as a marketing tactic. And when we think about where customer experience innovation is headed, it’s often personalisation (or the data needed to drive personalisation) at the centerpiece.

This makes it critical to think about your future state and plan backwards from that point, rather than simply making incremental improvements in the customer experience today. If your efforts aren’t a part of a larger plan to connect your channels and personalise the experience to drive long-term customer value, your priorities may need to be reprioritised.

In the end the decision of what to prioritise must come back to two things:

  1. Will the output allow you to better achieve your goals?
  2. Will the output give your consumers a better experience?

Personalisation often enables retailers to achieve both, so separating personalisation and CX improvements from each other may not be the right choice.


Q: How do you define the difference between segmentation and personalisation?

A: Short answer is that they are very different, but both are equally important.

Marketers segment customers based on demographics, historic behaviors, or predicted behaviors — such as women who have engaged in the last 30 days; or men from a specific region predicted to purchase in the next 7 days. Segmentation gets far more complex than these examples and oftentimes leverages multiple forms of data modeling and business intelligence teams.

Personalisation is different; it’s not about the group, but about creating an experience that’s unique for every individual. You can personalize without an incredibly robust segmentation strategy to increase engagement and results from email, web, and mobile, as well as with other areas of an organization.

Think of both segmentation and personalisation as strategies that have unique, but related, maturity curves. Take campaign personalisation for example. Within an email newsletter that goes to your segment of customers who have not purchased in the last 30 days, you might personalise recommendations to every individual in that segment based on recent additions to your product line or seasonal trends based on what that specific individual has historically browsed on your site and mobile app. Every individual in that segment gets a different mix of products, but how you personalise is based on your specific goal for the segment (in this case, re-engagement).

An advanced approach for that same segment of customers (who haven’t purchased in the last 30 days) you might personalise for channel, so that for customers who don’t respond to email, but who do respond to mobile push notifications, you send messages via mobile app and suppress them from email. There are a multitude of ways to combine segmentation and personalisation based on your specific goals.

A major caution here when evaluating technology partners is that many marketing technologies look to pass segmentation as personalisation — what you’ll end up doing here is sending a curated email to a specific segment, rather than truly personalising for the individual. Be sure you understand what’s truly possible before investing. You can find our more about how Sailthru personalises here.


Q: How do you analyse data to truly “know your customer” so that you can personalise?

A: There are two separate activities that go into this: 1) what data to collect and 2) how to analyse that data.

For what to collect, we believe that the best data sets to drive personalisation include a robust mix of demographic, behavioural, interest, and predictive data points. You want explicit data on who the customer is and where they are engaging from at any given point in time (geo-location). You also need data on how, when, and where they engage including channel preferences, purchase history, time of day, and more. Both explicit and implicit interests are critical to offering a relevant experience. Finally, the ability to predict what the individual will do next (open email? Purchase? opt-out?!?) will help you to personalise to promote (or detract from) specific outcomes.

In terms of analysis, it’s critical to move beyond campaign analytics and to instead take the long view by analysing customer cohorts so that you know what is most effective in delivering both short-term wins and in making long-term gains. An example of where this pays off is with one of our retail customers who started personalising message type. Instead of promoting product in every email, they engaged customers who were not predicted to purchase with content. If they’d been measuring on a campaign level they’d have stopped sending content emails very quickly, but because they’d been evaluating success at the cohort level they were able to recognise long-term lift in purchase frequency and average order value among those customers who received personalised nurture emails versus product promotion in every newsletter.