B2B Market Segmentation

The target market for any B2B product or service is not one homogeneous mass. Rather, it can usually be divided into several distinct groups based on who they are, how they behave, what they want or what they think.

The most common approach to segmentation in B2B markets is feature based. Here the target market is divided into groups based on ‘firmographics’ such as company size, location or activity. This approach may also segment the market according to the nature of relationship or product usage characteristics.

To segment based on features follow six steps:

1. Understand what features of a customer are relevant, i.e. will impact your proposition for or approach to the segment
2. Get a database of all your customers and record the relevant features next to each
3. Analyse your profiled customer base to identify the major groups which exist
4. Research competitors and speak to those receiving new business enquiries to additionally identify segments not currently in your customer base
5. Prioritise segments by mapping on sales and profit data from existing customers and accessing third party statistics on the structure of the business population
6. Identify which segments are likely to need a unique approach and detail how your proposition and messaging will reflect this

A feature based approach is useful, but it’s far more powerful to segment according to customer behaviours, needs and attitudes. This deeper level

of understanding lets you precisely tailor your offer and messaging so that you’re a perfect fit and the natural choice.

Use a four stage process to adopt this approach:

1. Begin by holding workshops with key customer facing staff to get their perspective on how customers differ
2. Then arrange a series of focus groups or in-depth interviews which represent a diverse mix of customers. These sessions will provide a vehicle to explore behaviours, needs and attitudes in detail
3. Next explore these areas in a more structured manner through a large scale quantitative survey. This refines the model by including a wider range of customers, allows each segment to be sized and most importantly it enables segments to be statistically derived
4. Then apply a Cluster Analysis. This statistic sifts through the data to identify groups of people answering questions in a similar way to each other but distinctly different to other groups. In doing so it accurately reveals and profiles the segments

This approach is a very different way of viewing customers for many and can be hard to grasp as it’s ‘softer’ than feature based models. So there’s a tendency for colleagues to intellectually accept the new segmentation, but eventually revert to their old ways. The trick to overcoming this tendency is to develop two plans – an internal communications plan and an activation plan.

Why Segment

The target market can usually be divided into several distinct groups based on:

1. Who they are - company features

2. What they do - buying or usage behaviours

3. What they want - needs, preferences and desired relationship 

4. What they think - attitudes

So the market for business networking events could be viewed as one market. However, this makes it very hard to develop an appealing event because not everyone wants the same thing. For example, some want a ‘party atmosphere’ but others want a more serious forum.

So whether you’re in the networking business or any other B2B business, a segmented approach will make your marketing much more effective:

•  A niche focus can be used as a solid basis for competitive differentiation
•  Products can be matched to (or developed for) the specific needs of each
• segment
•  Marketing messages (and channels) can be tailored to resonate with them
•  Lists of target customers can be bought
• The result is much higher ROI on any marketing activity because it tends to hit home first time. Sounds good, huh? Before we take a look at how you can adopt a segmented approach for your business, let’s tighten up the definition of what a segment actually is.

When a segment isn’t a segment

Any market can be segmented on any basis. If you wanted to, you could segment your customers into groups based on the number of syllables in their company name. You wouldn’t of course, but it illustrates the point that segments need to meet certain criteria. They need to be:

•  Meaningful – grouped on dimensions that influence how you’ll approach them
•  Distinct – each segment needs to be clearly different to others
•  Sizeable – each segment needs to be large enough to justify its existence
•  Identifiable – unless customers can be placed into a segment in ‘real life’, the exercise is useless
• If your background is in B2C marketing then it’s also important to recognise that B2B segmentation differs from B2C segmentation in several important ways:
•  There are usually fewer segments because B2B markets have far fewer buyers and these buyers tend to have similar requirements. This means there’s less room for diversity so a typical B2B segmentation will usually see four or five segments emerge
•  B2B markets tend to operate according to the 80/20 rule – 80% of revenue is generated by just 20% of customers. This means that a small number of very large customers should sometimes be viewed as individual segments in their own right. This is much less likely to be the case in consumer markets
•  The points above suggest that B2B segmentations are less complex, but they actually bring their own complexities. It’s sometimes hard to determine exactly what is being placed in a segment – an individual buyer, the buying team or the company they work for? And as B2B products themselves are often complex, teasing out the defining features of a segment can be equally complex
•  B2B segments tend to be based on ‘rational’ dimensions such as buying criteria rather than ‘emotional’ dimensions such as self-expression as is commonplace in consumer markets
•  B2B markets tend to move at a slower pace than B2C markets which means that segmentation models tend to remain accurate over the years, whereas in consumer contexts they can date quite quickly
• And finally, remember that no segmentation model is perfect:
•  There will be grey areas between segments where people display characteristics from both
•  Even within a segment, some cases will be more extreme examples than others
•  Not everybody can be categorised – some are just seemingly random or a law unto themselves

Feature based segmentation

The most common approach to segmentation in B2B markets is feature based. Here the target market is usually divided into groups based on ‘firmographics’ such as:

¢ Company size (either in terms of turnover or number of employees) ¢ Company location (country and/or region)

¢ Company activity or industry sector

Feature based segmentation may also segment the market according to the nature of the relationship. For example, existing customers may be grouped into categories based on value or frequency of purchase.

The major failing of this approach is that it doesn’t recognise that needs differ even within feature based segments. Admittedly, companies of similar sizes or in similar sectors do often share similar requirements, but there are usually important differences between them. For example, many technology companies segment their market into SMB (small and medium sized businesses), Enterprise (larger businesses) and Public Sector. Let’s take one of these segments to illustrate the point. SMBs, especially in relation to technology, differ wildly in their behaviours, attitudes and needs. Some are luddites; others are at the cutting edge. Some have a dedicated IT function; others rely on third party expertise. Some view IT as mission critical; others could temporarily survive without. I could go on, but you get the point. So to treat all SMB customers as the same misses a big opportunity.

That said there are some compelling benefits to adopting a feature based segmentation:

•  It does reflect differing customer needs to some degree
•  It’s easy to implement – identifying segments and assigning customers is relatively straightforward
•  It’s actionable – people ‘get it’, lists of target customers can be bought and marketing channels identified (e.g. industry trade media)

For these reasons, feature based segmentation is heavily used in B2B markets. You can do so by following five steps:

1. Understand what features of a customer are relevant, i.e. the features which will impact how you should approach them. Often talking with those on the front line (sales teams, customer service, etc.) provides good guidance in this respect, and usually company size and sector emerge as the two primary features

2. Get a database of all your customers (easier said than done perhaps!) and record the relevant features next to each. When doing so use clear cut categorisations so that you’re able to analyse the database. You can create these categories however you wish, but commonly used size bandings are sole traders, 1 – 9 employees (also known as micro businesses), 10 – 49 employees (small businesses), 50 – 249 employees (medium businesses) and 250+ employees (large businesses). Industry sectors are typically categorised according to SIC codes – a series of four digit use a feature based codes many government statistical bodies use to classify industries. The benefit of adopting one of these widely used methods is that when looking to buy in prospect data or size the market using third party statistics, your segments can be directly matched to those of most data providers

3. With a profiled database in hand, analyse your customer base to identify the major groups which exist. But don’t stop here. It’s quite possible that there are segments in the market who are not currently your customers. To identify these take two steps. First, speak to those receiving new business enquiries to understand if there are prospects whose ‘type’ is under-represented in your customer base. Second, research competitors to see how they segment their market (often their website will reveal this or if they are publicly listed, investor strategy presentations can be a gold-mine of information)

4. Now that the major segments in the market have been identified you need to prioritise them. Mapping on sales and profit data from existing customers provides good guidance in this respect. You may also find it useful to access government statistics on the structure of the business population as this can reveal the total market and segment size

5. And lastly, identify which segments are likely to need a unique approach and detail how your offer and messaging will reflect this

Behavioural, needs and attitudinal segmentation

A feature based approach is useful, but it’s far more powerful to segment according to customer behaviours, needs and attitudes. This deeper level of understanding lets you precisely tailor your offer and messaging so that you’re a perfect fit.

Like all good things though, this approach is not easy. Effort needs to be invested into fully understanding customers and developing a comprehensive segmentation model. The resulting segmentation also needs to be carefully implemented if it is to gain traction.

It is worth the investment though, and the ideal is to develop a segmentation model using a four stage process:

1. Begin by holding workshops with key customer facing staff to get their perspective on how customers differ. Starting with this internal view will help to design subsequent stages of the research and guide the line of questioning followed. These sessions should use interactive, thought-provoking exercises to create an initial hypothesis of how many segments there are in the market and what makes each one different from the next. For example, workshop delegates could be asked to stereotype the different customer types they’ve met and act out a scene involving them
2. The internal view is a good starting point, but it’s essential that any segmentation model is based on insights gleaned directly from the customer. Steps two and three ensure this is the case. Arrange a series of focus groups or in-depth interviews (which depends on the situation, e.g. if customers are geographically dispersed or fierce competitors then focus groups aren’t appropriate) which represent a diverse mix of customers so that all possible segments will emerge. These sessions will provide a vehicle to explore behaviours, needs and attitudes in detail. After a while you’ll begin to notice distinct segments emerge in relation to each of these dimensions. This provides an initial segmentation model. The dimensions this model is based on will be unique to your situation, but common factors which define different B2B segments include:

3. The qualitative approach adopted in stage two is great for creating an initial model, but a more structured approach is needed to verify, size and profile segments. So stage three does just that. It explores behaviours, needs and attitudes but this time does so in a more structured manner and with a larger group of the target market (e.g. through rating scales). There are three reasons for doing so: it refines the model by including a wider range of customers, it allows each segment to be sized and profiled, and most importantly it enables segments to be statistically derived. In this step also be sure to include questions around their spending behaviours and marketing channel preferences. This lets you prioritise high value segments and tailor your marketing approach

4. Step four is the science bit. Armed with lots of data from a representative sample of customers, you need to make sense of it and reveal the segments.

To do so have a statistician apply a Cluster Analysis (Cluster Analysis is actually a family of different statistical techniques which includes our preference, the Cluster Ensemble approach). In layman’s terms, Cluster Analysis sifts through the data to identify groups of people answering questions in a similar way to each other but distinctly different to other groups. In doing so it accurately reveals the segments. And more than that. It sizes each segment by identifying the percentage of those interviewed falling into each group, it profiles each segment by overlaying firmographics and, if you’ve included sales data in the analysis, it can even identify the most profitable segments. Overlaying this information is invaluable when it comes to deciding which segments to prioritise and actually targeting them (e.g. if a priority segment is most prevalent in a particular sector you can focus marcomms on relevant trade publications). As well as identifying segments, Cluster Analysis also identifies a small number of ‘magic questions’ from the survey.

These questions are the most powerful in assigning someone to the right segment, so by asking them in ‘real-life’ you can quickly assign anycustomer to their segment and act accordingly. This capability comes into its own when implementing the segmentation (see later)

Now you have a clear map of your target market. The different segments have been revealed, their needs have been identified, they’ve been prioritised based on size/value and you can find them in real life as they’ve been profiled firmographically.

Implementing the segmentation model

There’s one final step though. You have a carefully crafted segmentation, now you need to use it. This is harder than it might sound and countless segmentation exercises fall at this last hurdle.

The problem is one of human nature. A behavioural, needs based or attitudinal segmentation is a very different way of viewing customers for many. It can also be hard to grasp as it’s ‘softer’ when compared to feature based models. What’s more, a ‘soft’ approach like this can be harder to use on a day-to-day basis. You may have a proposition perfectly matched to the needs of segment X, but how can a sales person contact them? Or if a prospective customer makes an enquiry, how does the sales person identify which attitudinal segment they are in so that the right ‘hot buttons’ can be pushed? So there’s a tendency for colleagues to intellectually accept the new segmentation, but revert to their old ways soon after it’s introduced.

The trick to overcoming this tendency is to develop two plans – an internal communications plan and an activation plan.

The internal communications plan is designed to spread the new view of the customer throughout the organisation and ensure that each segment is fully understood. Every plan is unique but it’s important to:

• Bring segments to life by giving them a memorable name and creating a ‘pen portrait’ describing them. Let your creativity loose here for maximum impact. For example, use actors to create a ‘meet the segments’ video or place life size cut- outs around the office
• Involve all customer facing staff as they’re the ones who’ll need to use the segmentation. ‘Involve’ is the active word here. Colleagues will betterunderstand the segments if they need to actively learn about them. Again, creativity pays dividends here. For example, role-playing exercises give aperspective on the segment which words can’t match

But equipping colleagues with an understanding of and belief in the segments is not enough. You need to provide them with the tools to act on the knowledge – you need an activation plan. There are three parts to this plan:

• The ‘marketing’ section details how your offer and messaging strategy will differ by segment. Think about their unique product needs, how they want to interact with you, what will keep them loyal (or make you first choice if they’re a prospect) and what communication channels they prefer
• The ‘relationship’ section is designed for sales and account management teams. It provides them with easy to use tools to identify which segment a customer falls into (see ‘magic questions’ earlier) and a ‘cheat-sheet’ outlining how they should tailor their approach in light of this. It also provides details of which ‘firmographic’ categories each segment is most prevalent in so that target lists can be bought
• The ‘systems’ section contains a plan to embed a segmented view. The most important action here is to assign each customer on your CRM database to a segment as this allows marketing campaigns to be tailored. This can be achieved by asking ‘magic questions’ to all new customers and to existing customers whenever the next contact is made. Of course, this won’t always be practical so an alternative is to model the data. This approach begins by assigning a sample of customers to the correct segment by asking the ‘magic questions’. It then applies statistical analysis to uncover other characteristics which make members of a segment identifiable and are already held in relation to all customers, e.g. most members of segment A have bought product X and contacted customer support three times in the last six months. As the characteristic data is already held for all customers, it can then be used to assign them with a reasonable degree of accuracy to the right segment.