How to segment your users
Neglecting niche customer segments can cost you market share. Why is user segmentation so critical? Why do most segmentation frameworks fail? How can you get it right?
When I started at OLX (Avito), our strongest competition was niche players.
Those companies captured a chunk of our market by targeting an overlooked user segment. They offered better solutions, understood the segment’s needs better, and delivered a stronger value proposition.
As a classifieds platform, we’ve been a successful product of mass appeal. You could find and arrange deals on almost anything from iPhones to Cars and Real Estate. However, this mass adoption was largely due to a first mover advantage.
Users flocked to the service in the absence of alternatives. Once a better alternative appears, churn is inevitable. For instance, one player over-invested in a superior car catalogue. This slowed our growth. To counteract this, we adopted a strategy of building vertical experiences ourselves, both operational and product-led.
Segment-driven disruptive innovation is a key risk for any large business. Identifying segments, understanding their pain points, and crafting a value prop is essential to continuous growth.
Let’s dive in and understand why this disruption happens, how to segment your users, and what are the key pitfalls to avoid?
📢 Products of “mass appeal”
Understand your customer. Pinpoint their key problems. Add value.
Those are clichés the product theories are flooded with.
Most lean startup and product literature claims you cannot succeed without customer knowledge. However, with some luck and happenstance, it is very possible. Over the past 10 years, I’ve worked at highly successful companies, all valued in billions, that had little customer knowledge. My job was to reverse engineer why the product was used and improve the decision-making framework.
🚀 How to launch a “mass appeal” product
The common pattern across all of those companies was that they were building a “mass appeal product” that could be used pretty much by anyone. However, for such a product to succeed, multiple things need to fall in place.
First, the blue ocean market context. The company needs to be a first-mover without direct competition and with complex, inefficient alternatives.
Second, the opportunity needs to be large. Ride-hailing, food deliveries, or any repeatable activities are examples.
Third, the product needs a solid sales-led/product-led go-to-market strategy.
🛡️ Vulnerability to niche competitors
Despite hitting the growth jackpot, such “mass appeal companies” are very vulnerable to niche competitors.
Take Adobe Photoshop. First launched in 1990, it targeted a wide range of user segments. Photographers, illustrators, graphic designers, brand designers, and advertising agencies were all in the mix. There was a period when Photoshop was the clunky go-to for product designers.
However, over the years niche-focused alternatives popped up. Figma ate up the product designer segment (alone valued at $12.5b), Canva took over non-professional illustrators (valued at $26b), GIMP bit off amateur photographers, and Procreate captured illustrators working on an iPad.
This happened because Photoshop offered a mass appeal product with suboptimal value for specific user segments.
🚫 How not to segment your users
❌ Personas
I’ve seen a team investing 6 months of research to identify their users. The output was 20 personas with unique journey maps for each.
Jennifer, 34 years old, a mother of two, is an active middle manager, who wants to save money for her kids tuition and retirement
Josh, 22 years old, a DJ, works part-time as a bar waiter, spends all his money partying
The findings were presented in a slide deck and highlighted a few curious pain points. However, after the meeting, the deck was archived and never opened again.
The team struggled to make sense of the insights. Should they double down and build a product for mothers of two? DJs? Or a variety of other equally important segments? The insights were not actionable enough.
Because personas require very specific user attributes, the results are often ambiguous. You can create an endless number of personas, some of which may even be contradictory. However, all these personas represent active users of your product.
❌ Jobs to be done (JTBD)
JTBD was a pretty hyped framework a few years back. There were even multiple schools of thought on the subject (you can check out Clayton Christensen’s and Alan Klement’s works on the topic).
Essentially, it doesn’t matter what persona your user has; what matters is the job they need to solve. Both a mother of two and a young DJ need a hammer to hang a picture in their living room.
On the surface, it seems like a reasonable alternative to personas. However, in practice, JTBD rarely yields actionable insights.
For instance, you can structure trips by the type of job: commuting, shopping, leisure, weekend getaway, etc. However, this approach ignores context. What about price sensitivity? (e.g., a student is more price-sensitive than an office worker). What about commute patterns? (e.g., going to the university differs from visiting a workplace).
While personas are overly detailed, JTBD leaves out too much important information. Over the past 10 years, I have not seen a successful and continuous implementation of this framework.
❌ Socio-demographic
Age, social class, gender, and occupation are all great for sizing up the market, but almost completely useless when it comes to user segmentation.
Let’s say your average user was aged 25-34 years, was a representative of a middle-class tech manager segment, and worked as a coder. How actionable would those insights be?
❌ Geography
Same applies to regions, countries, or cities. Yes, there might be a correlation between a country’s GDP per capita and average price sensitivity, but that’s pretty much it.
Alternatively, going too granular on the topic also leads to ambiguity. Are you optimizing the value proposition for your user or for an h3 polygon cell with a res of 2?
❌ Cultures
Almost 10 years back, while researching their “Next billion users” strategy, Google did dive deep into the cultures.
For instance, the perception of time and space between Western and Eastern populations is different. This reflects in the usability patterns of apps and services.
Those insights, while thought-provoking, are pretty much useless for product teams working on their local problems.
🗂️ How to segment your users
My key message is that there’s no single framework. Existing frameworks often cause analysis paralysis by being either too granular or not granular enough.
What my team and I have discovered is that principles work best for customer segmentation. The rest relies on your creativity.
📏 Key principles for user segmentation
MECE
Taken together, those segments need to be mutually exclusive (unlike Personas) and collectively exhaustive.
Actionable
Deep enough, but not too deep
The level of granularity of user attributes description needs to be deep enough, but not too deep. A middle-aged office worker is deep enough, a mother of two is already too deep.
Wide enough, but not too wide
Thinking in region or city-wide in most cases is okay. Country-wide can also work, but region (e.g. EU) already leads to the loss of context.
User and not intent focused
Focus on intent leads to the loss of context. It is key to focus on the user and her attributes instead of the job that needs to be solved.
Simple
It should be simple to understand and memorize.
✅ Good example (for a mobility service):
Office commuters (25-34 yo)
Travels twice a day, from home to work and back at typical commute hours. Medium price sensitivity.
Students (18-23 yo)
Travels twice a day to the Uni, once early morning, once early evening. High price sensitivity.
Tourists (18-35 yo)
Visits a city on a one-two day trip, interested in sightseeing. Low price sensitivity.
📝 TL;DR
✅ Products of mass appeal and their vulnerability:
Products of mass appeal are businesses designed for everybody without a clear understanding of their users.
It is possible to launch such a company if it’s a blue ocean market, large opportunity, and there’s a strong sales-led/product-led GTM strategy.
Such mass appeal companies are very vulnerable to niche competition.
✅ How not to segment your users:
Personas lead to ambiguity and analysis paralysis.
Jobs-to-be-done leave out too much important user context.
Socio-demographic segmentation is too high level.
Geographic segmentation is too high level or too deep.
Cultural segmentation has purely theoretical interest but little practical applicability.
✅ How to segment your users:
There’s no specific framework; all existing ones are purely directional. They either lead to analysis paralysis or insufficient depth.
It’s best to rely on principles to segment your users.
Key principles are: MECE (mutually exclusive and collectively exhaustive), actionable (deep enough, but not too deep / wide enough, but not too wide), user and intent focused, simple.