Segment maps: The view of your customers you didn’t know you craved

So, you’re trying to make your marketing spend work harder than ever in the middle of a cost crisis? You’re trying to retain that precious customer base at a time when they’re more likely than ever to jump ship? Your boss wanted a plan yesterday and blanket discount codes just aren’t going to cut it this time?

Fear not — because it’s time to start working smarter, not harder!

For years we’ve thought of ‘omnichannel strategies’ and ‘hyper-personalization’ as buzzwords, but it’s now time to invest, give your business the self care it deserves and keep ahead of the curve. Research shows marketers using three or more channels in any one campaign earned a 287% higher purchase rate than those using a single-channel campaign (Omnisend, 2020). Sounds like a no-brainer then, right?!

We know customers have higher expectations than ever when it comes to online (and offline) purchasing experiences, and in the current economic climate businesses need to work smarter to give them the experiences they want and, let’s face it, have come to expect.

As part of Peak’s Customer Intelligence stack, the Audiences application is designed to make it easy for marketing teams to create, understand and action valuable segments to enable a truly personalized marketing strategy. It’s been built to complement and enhance existing martech solutions, in no way replacing them but instead centralizing that data and adding intelligence.

This, in turn, drives key acquisition and retention metrics, ultimately generating revenue growth and profit for the business. Our aim is to support businesses on their journey to provide customers (or potential customers) with a unique experience that is tailored solely to them; some may call this a one-to-one personalization strategy.

One thing to note is that it doesn’t just stop at segmentation; a successful personalization strategy equation would take into account the who, what, when and where. This could be seen as a journey-centric strategy rather than just a segment-centric strategy. Let’s delve into this a little bit further and start by explaining what we mean by segment creation.

Who?

To start slicing and dicing your customers into actionable cohorts, it’s important to first understand your customers by capturing their behaviors and preferences. Simple, right…? 

The real game-changer at this stage is using the power of artificial intelligence (AI) to start predicting their next move (think an artificial Anya Taylor-Joy in The Queen’s Gambit!) 

This will ensure that the content being served is both relevant and timely. The two ultimate crimes in the marketing world are to waste money and to lose customers, so capturing and leveraging data is your secret weapon in avoiding these. No one wants to spend money trying to acquire people that don’t want to buy, so applying an in-market model can increase return on investment (ROI) significantly, often with immediate effect.

Coupled with something like a churn prediction model, you can also identify those who are about to go elsewhere; so this is your opportunity, and cue, to pull out all the stops to save them and protect your customer base. Depending on the quality and granularity of your data, the sky really could be the limit in this space.

When?

These models also start to touch on the when. We know that bombarding customers with everything all the time is damaging. A quick click of the unsubscribe button and your brand is gone from the forefront of their minds forever — or at least for a short while while you work extra hard to win them back (at an extortionate acquisition cost!) So, ensuring that you’re visible in the right place at the right time is key, and using AI to help define frequency and recency strategies will help you win.

This is all well and good, but how do we actually find and create the right segments, with so many data points to choose from?

Segment Maps is a new feature pioneered by the Peak team as part of the Audiences application. It allows you to visually explore and leverage predictive and non-predictive customer attributes with ease. By selecting important attributes and narrowing down your cohort, using the filtering functionality, you instantly have a granular but digestible breakdown of your customers, along with key metrics to aid the decision-making process. This view allows you to create well-informed segments that can then become the base for your cross-channel campaigns.

Makes sense! So, what’s next?

What?

You’ve now identified who you want to market to and may have used predictive attributes to help determine the when. It’s now time to understand where and what to speak to your customers about! You may have created the segment with at least one of these questions in mind, but it’s important to have an understanding of all of the who, when, what and where to maximize the results of your campaign.

Understanding past purchasing trends definitely helps with creating relevant content, but this is where AI can be leveraged again to help predict the preferences and behaviors of your cohort. A perfect example of this is pairing the Audiences application with Peak’s Recommender application to supercharge the customer experience with user-specific product and category recommendations.

This takes historical customer and product metadata into account to predict next purchase preferences, giving the content creators the information they need to create relevant assets for that segment and even embed user-specific recommendations into their communications. In addition to making it infinitely easier to decide what content to run, understanding preferences for customer segments could also help inform the bidding strategy for channels like Google Search and PLA. This means that not only are we providing more relevant content to improve conversion, we also have the ability to optimize costs at the same time.

Where?

The majority of businesses these days do prioritize personalization as a strategy to an extent, but it often falls short when these initiatives are siloed across the different touch points of a customer journey. It’s understandable why this happens, though; teams want to be autonomous and own their own systems to make their jobs easier, optimizing their specific team’s KPIs. Business leaders also want to empower individual teams to move fast and unlock their potential. 

As businesses grow it becomes harder and harder to unpick this behavior and align systems, data and teams (even if you’re lucky enough to have a huge tech department to support this.) Time and time again businesses pump endless resources, time and money into resolving this at the core and never seem to come out the other side (my thoughts are with those still lost in the code!)

It’s not all doom and gloom, though, because there is a way forward! Working backwards from a small use case supported by a centralized AI platform allows you to tackle the challenges and unlock value incrementally. The more challenges unlocked, the more the synergy effect will start to kick in. 

Circling back to our segments, we may want to integrate this with one, two, three or even more systems so we can ensure consistent messaging across the key touch points for those customers. Our concept of ‘headless segmentation’ at Peak tackles just this. We can also start to understand channel preferences to set a strategy on where to go hard and where to ease up on spend for certain groups, further optimizing spend and improving the customer experience.

Great…so, is that a wrap then?

Well, it’s a good start! We have the tools we need in Audiences to identify who, what, when and where, so it’s time to get some of these campaigns out in the wild. But what does success actually look like? It depends on who’s asking!

It’s all about the value

I’ve been there — it’s been ten minutes since a campaign went live, and people are eager to know how many millions it’s generated, especially when it’s a new and exciting concept or strategy. In my experience, it’s all about setting expectations, baselines and micro-metrics. Defining and agreeing the key goal up front alongside calculating the baseline metrics is a great start — and it’s even better if we can define control groups for A/B testing! 

Another big challenge is to get an early reading of success, or the ability to meet statistical significance on the key metric, with all the other business initiatives adding noise. This is where micro-metrics help to provide indicative results, and these work best on metrics that are isolated and higher in frequency. 

For example, think of clicks as a micro-metric of orders. Clicks happen much more frequently than orders, so if clicks are going in the right direction, then it’s an early indicator of success (or at least enough to reassure you that you don’t need to pull the plug on this just yet!) It’s also likely that there is room for improvement after the first pass; this is just the start, and running a series of test-and-learn experiments will move the dial quickly. 

Value is the reason why we’re all in this game, so it’s really important to keep the focus on this from the very start to the end…or ideally through to the next iteration. Our AI-powered Audiences application is there to provide the tools you need to get to this value and optimize on the results. It’s a really exciting time to be working in this area; we’re able to use data like never before, faster than ever before — so just remember to enjoy the journey!