If you can answer three simple questions, you have damn good market segmentation
Effective market segmentation is the most reliable and effective means of answering the following three questions:
- Who are our customers?
- What do they want?
- How do we better engage with them?
Being able to answer these three questions is a cheat sheet for businesses. Why? Because an understanding of your customer keeps your business relevant.
Market segmentation gives businesses the foundations to maintain relevant and valuable to their customers at a reliably scalable level. This article introduces to a modern way of crafting an effective and practical segmentation strategy (one that we use at Tiny CX).
It is not a wholly DIY solution (for most), nor is it not a lean solution. It is, however, the solution that can reliably help you answer the three golden questions.
1. Who are our customers? The role of data
The amount of data in the world is staggering. By 2025, it’s estimated that, globally, 463 exabytes of data will be created every day; that's 40x more points of data in the world then there are stars in the observable universe. How can you harness that?
No company has a perfect data strategy, in fact, most companies have significant gaps in their data. That, however, doesn't inhibit you from painting a significant picture as to who your customers are. You can do effective segmentation with just a little bit of data.
Use cluster analysis to inform segmentation strategy
Your first question, "who are our customers?", is solved by data. A data scientist can do a preliminary cluster analysis of your database to create an initial set of hypotheses for your segmentation strategy. You should be looking to create clusters/segments that have the following traits:
- Easily identifiable: there is no point using segment variables that are hard to find;
- Mutually exclusive: ensure that you can easily separate different clusters - our rule of thumb is "can we confidentially assign a cluster/segment to this customer with 6 data fields/questions or less?"
- Value-based: It's all very well segmenting by age, gender, and location if it's relevant, but you should ideally be utilising elements that are relevant to your business. If you can, use variables like, "how much do they spend?", and "do they pay on time?" to help distinguish your segments;
The best news? You can achieve this with a relatively incomplete database and someone who knows how to do basic cluster analysis.
Surveys: The 2nd-best option for segmentation
Surveys were once the most-effective means of answering "who are our customers?", but big data has surpassed them in terms of quality and quantity of data. Be careful when using surveys for segmentation strategies: the results simply aren't as reliable as what big data can tell you.
2. What do our customers want? The role of research
What do your customers value when it comes to the products and services that you give them? The key to knowing what your customers want is as-follows: know how their perceptions are formed.
Daniel Kahneman and Amos Tversky famously discovered the concept of "systematic human cognitive bias", or how our perceptions and emotions drive our rational thought process. In their book, "Thinking, Fast and Slow", they show how people have two types of systems when it comes to making decisions:
- System 1: Our "animal instincts" that are largely emotion-driven, which in-turn inform;
- System 2: Our rational mind
I won't regurgitate the theory to you (here is the book), but will summarise one of the implications from the discovery: our "logical" processes are systematically biased by our emotions and perceptions.
What does this mean for companies? You can't understand people easily. They are irrational and highly emotional, and as such, you need to study them within the context of your industry, category, and offering to learn how their System 1 informs their System 2. This cannot yet be done effectively by big data or surveys.
Instead, you will need to embark on a series of interviews with members of your identified customer segments to identify what makes up the concept of "value" for them. By asking "why do you value the way you do", you are breaking down how System 1 is influencing System 2. Doing it within the context of segmentation will enable you to understand the common themes that distinguish and/or unite segments.
How to Interview People: Contextual Interviews
The reader who knows how to interview can skip this section and continue onto Part 3: How do we better engage with them? How to apply market segmentation.
A simple technique to use is through contextual interviews, which provides an interviewee with stimuli (related to your offering) that they can use to frame a response. Note that this each interviewee must be tagged against the outputs from your initial market segmentation - you're wanting to know what sets each segment apart.A contextual interview example is provided below. Note how the interviewer is trying to find one thing: how system 1 is influencing system 2.
*Interviewer shows a picture of a product*
Customer, "I don't like that style"
Interviewer, "Why not?"
Customer, "Because it doesn't respect my time. It's a great product, but it's so slow to use and I would prefer a faster solution over a more enjoyable solution.
" Interviewer, "So you value speed over enjoyment?"
Customer, "Because I want a company that respects my time. For me, if they show respect for time instead of trying to 'wow' me with their products, then I'm more trusting of their intentions"
Customer, "So you don't like this style because it doesn't respect your time?"
Insight: The product has to be built so that it is efficient to use because the customer values their time and trusts companies who respect that.
For a demo of how we do this at Tiny CX, you can view our 10-minute demo video here.
To summarise, the interviewer had a clear goal: to identify the customers' value drivers within the context of the product. They did this by prompting system 1 (emotional response) and understanding how it connects to system 2 (logical).
Part 3: How do we better engage with them? How to apply market segmentation
My friend Will Anderson from Freeform Strategy puts it best, "research doesn't stop until it has been applied". At the end of months - and sometimes years - of work, market segmentation strategies can often be printed, bound, and put on the CMO's shelf to collect dust. Why does this happen?
Market segmentation, theory or practice?
The consequences of getting segmentation wrong can be severe - just ask any CMO that has spent millions on a campaign that flopped due to poor targeting or irrelevant messaging. To avoid getting it wrong, companies often call in research specialists: people who are good at the methodology of research but are horrible at the application of research.
This may not seem like a significant problem until you receive the final research report from your chosen specialist. A criticism that I too am guilty of being on the receiving end of goes something like this,
"This segmentation strategy is fantastic, but how do we apply it?"
The reality is that effective segmentation needs two things:
- Rigour (the typical strength of the researcher); and
- Practicality (a typical weakness of the researcher)
Having both skills is a must. You cannot get by with great research that cannot be applied, just as you cannot apply a segmentation strategy that is based on unreliable research. The use of "big data" and pursuit of qualitative insights will address your need for rigour when done correctly, so now we turn to requirement 2: practicality.
First, I would like to acknowledge that there is no "one way" to apply research. There are, however, key outcomes that you need to achieve. This will outline the key outcomes that mark a successful market segmentation strategy rollout plan.
No matter how much data and research you do, there is the need to validate your segments before you roll out a scaled strategy. This is an often overlooked phase of a market segmentation project. You can be creative as to how you validate, but here are two strategies that work for me:
1.1. Validation via survey
It's old-school, and there is still a chance that there will be a difference between what people say and what they do, so it isn't perfect. It is, however, a straightforward path to validating your segments and their insights. Simply identify people within your segments and send them a survey to measure how strongly they agree or disagree with the insights that you have identified. If there is not a strong correlation, there may be a problem. The nature of insights is that they aren't always known to the customer. Take the example of the original "shake and bake" cake mix. When it was first tested, people said they loved it, but when it came time to buy them, sales never took off. Why? Because people felt guilty, they felt like "just adding water" was cheating. In the research, people were fans, but behaviourally, they baulked - it took validation by behaviour to fine-tune the product (by introducing the need to add an egg, and thus reducing the feeling of guilt).This is why validation via behaviour is better...
1.2. Validation via behaviour (preferred option)
Validation via "test and learn", whilst slower, is a method well-worth making the time for in your segmentation strategy, as it enables you to see customers behaving according to their segments and value drivers in the flesh. This has two benefits:
- It's reliable - you're seeing the proof in the putting;
- It makes it easier to understand how to apply the insights by witnessing how they manifest.
Example: validating via Contextual Inquiry
Contextual Inquiry is, in layman's terms, watching people's behaviour and occasionally interrupting them to ask why they're behaving that way. The added bonus of CI is that if you video it, it becomes a useful internal communications asset (more on that later).For example, let's say I undertake a process of contextual inquiry to observe a segment of middle-aged stockbrokers [the customer] having new suits fitted by a tailor [the business]. If research tells me that this segment is driven by the insight, "I don't know what I want, but I know what I like", then I can better understand why a customer I am observing keeps deferring to the tailor for suggestions about colour, style, and cut. In fact, my observation could enable me to split the insight into two in order to make it easier to apply. For example, one split could be, "I know what I like, but I don't know what that is" and the other could be "I need the tailor to recommend something in order for me to feel comfortable in the decision".
Finally, in this instance, CI would probably even give me the ability to start figuring out how to optimise the tailor experience for that segment then and there. For example, I could have the tailor bring out a style/look book so that they can profile the customers' preferences and assertively recommend potential cuts, colours, and styles.
Note: In the instance where an insight does not match up with behaviour, be sure to test within a different context. Some insights may apply only to certain stages of the customer journey, and it's worth your while testing insights against each stage of the journey. E.g. "Don't sell to me" as an insight might only apply during the early stages of the customer journey - when it comes time to buy, customers may actually like being sold to!
2. Internal communications
How you present your new segmentation strategy to your internal stakeholders can influence how effective it is. This may seem contentious and oversimplistic, but I have a single heuristic that I follow when determining how to communicate changes in segmentation whether I am communicating upstream or downstream.
Upstream: Executive and Board
Be sure to communicate your segmentation strategy to your executive team and board in terms of the impact it will have on the value of your organisation. Key benefits of proper segmentation are:
- Increased marketing performance
- Increased operational performance
- Decreased churn
- Increased customer value
- Increased marketing efficiency
- Increased operational efficiency
If you can quantify the segmentation strategy in those terms, my experience then shows that executives and boards then simply want to know what those segments are and what their profiles are at a high level, as well as the implications for the rollout phase. For this purpose, I generally prefer to couple a report with a video (preferably taken during Contextual Inquiry) to make the most use of the limited time I have.
Downstream: Management through to frontline
This is where it becomes a matter of practicality, "How do I apply this segmentation strategy?" is the #1 question that you will be hearing from this group.There are SO many ways that you can brief your segmentation strategy to managers and the frontline, but the structure that I find works best is as follows:
- Big picture: Ensure the staff see what value the segmentation strategy will bring to the organisation (you can use the "brief up" assets for this);
- Context: Breaking it down into what it means for each department is going to help. Using the Contextual Application rules (step 3) will help, but the important thing is to make sure you communicate in a way that is relevant to that department and/or role. For example, tell your operations how the segmentation strategy will require systems and processes to be more tailored to specific segments and insights.
- What needs to change: This is the big one, and depending on the size of your organisation, you may need an audit to help you figure this out. Segmentation strategies inevitably change the way that businesses are run, which means that the staff member and department need to know what they have to do in order to update their BAU to support this new change.
3. Contextual application & operationalisation
It is during this stage that many companies stuff up their segmentation strategy. Often, the marketing department will nail the application of their segmentation strategy (because it's their department that is responsible for it), but in operations, sales (if they're siloed from marketing), and *gulp* the inexplicably siloed "CX" department, they're so much as giving the segmentation strategy a passing glance and asking, "so what?".
As the owner of a segmentation strategy, it is your responsibility to ensure that every department has the tools and the training to be able to implement the strategy effectively. This isn't hard to do - it's about consulting with your other department heads to (a) audit where change needs to occur and (b) identify the most effective means of re-training staff so that they can use the segmentation strategy in their strategies and BAU. When planning your company's segmentation, look forward to this stage in the process and allocate time and resources to ensure that you are including the process of "contextual application". Below is a checklist that can be useful in determining how you can liaise with your other departments to ensure that CA is properly considered.
Contextual Application planning checklist
- What part/s of the customer journey does this department have an impact on?
- What processes, systems, & assets (e.g. customer portal) will need to be updated as a part of the new segmentation strategy?
- How much will it cost to change the processes, systems, & assets?
- How are staff currently trained in new processes?
- Who needs to be retrained as a part of the new segmentation strategy?
- How can you best re-train staff?
- How many hours of training will be required? Who will do the training?
By asking these 7 questions, you should be able to create a comprehensive and exhaustive segmentation rollout plan that contextually applies and operationalises your new segmentation strategy.
4. Measurement & optimisation
This is the Rosetta Stone of segmentation: if you can measure its impact, then you can understand how to optimise it. There are a lot of different ways of measuring your impact, and if I'm honest, I'm not enough of an expert in above the line metrics to be an authority on the matter (there is plenty to Google, if your heart so desires).
Instead, I'm going to introduce you to a highly practical approach that I have called "The Hourglass" methodology. It is a very useful heuristic that can be used by anyone to measure and optimise the impact that their segmentation strategy is having on their customer acquisition, retention, and value. Here's how it works:
4.1 Map out your customer journey
Mapping out your customer journey is an exercise in auditing all of your brand touchpoints. The result of this process gives you clarity and scope as to what needs to change.
I strongly suggest using a granular framework (our framework, the AICEDS OSLA, can be found here) as a reference for creating your own journey, as the importance of nuance and detail can never be underestimated in journey mapping (seriously, if you're weak at this, spend $48 to learn more here). The traditional "AIDA" model runs the risk of oversimplifying things - be careful in using it!
4.2 Map assets to the journey
An "asset" is anything that serves to "nudge" your customer from one stage of the journey to the next. It could be a sales script, a retail store, an ad, a website (or, more likely, a page on a website), an app, or a moment within a service design blueprint. You should have done the bulk of this work within the "contextual enquiry" phase, so don't freak out - you simply need to piece it all together.
4.3 Define how each asset is relevant to your segmentation strategy
Not every asset is for every customer, nor is every asset designed to deliver on every insight. Ensure you (a) assign which customer/s an asset is intended for and (b) what insights are being delivered on via the use of this asset and/or channel.
4.4 Assign a "goal" to each asset within the journey
80% of the time, it's conversion rate. A simple equation of "how many arrived at this asset?" vs. "how many progressed to the next stage of this journey via this asset?". You may also want to consider churn rate, cost of conversion, referral, DAU, MAU, time at asset, or another form of engagement measurement.
Once you know what you're measuring, it should be pretty simple to figure out how to do it. For small-scale companies, spreadsheets are generally sufficient. For scaled/scaling companies, solutions like Hrglass, Funnelytics, Hubspot, and Salesforce can be handy. Outsourcing your data management can be really useful as well, and the folks over at Athyna can be highly helpful in providing Virtual Assistants to assist in this regard.
Once you can measure, you can optimise. Again, there is a lot of literature on optimising, from simple A/B split tests to more advanced API-based solutions like Segment, Optimizely, and MixPanel. We are creating a new course on using data to optimise your marketing (for SMEs), so if you're interested, make sure you check out the CX Academy.
Conclusion: This is not an article...
...This is intended as a manual. As such, please don't feel overwhelmed by the information within. Rather, come back to it and try to apply it to your process. Better yet, iterate on it - if you have an effective way of improving on it, do it and tell me about it!And finally, tell me how you go with it!