Trust Me, I'm a Marketer: Segmentation

A note before you read. This is the first in a short series I wrote around 2010, called “Trust Me, I'm a Marketer.” I'm republishing it as a foundation piece, because it's where a line of thinking I'm still developing began. One honest caveat: the mechanics here use a segment as a filter - deciding who to mail and who to ignore. I'd frame it differently today. A segment is best treated as one input to understanding a person, not the output that sorts them into a mailing. But the principle underneath held up, and you'll see its first form in the warning at the end of the second piece: protect the audience that trusts you. These days I'd build that at the level of the individual patron, not the segment.

Welcome to the first in a short series I called “Trust Me, I'm a Marketer” - because that, in the end, is the job: earning the trust of your audience. We start with the foundation of everything that follows, segmentation. This piece covers what it is and a step-by-step way to do it yourself. Later pieces look at how to use it in your communications, and the strategies that build that all-important trust.

Segmentation

For 20+ years arts organisations have gathered customer data at the point of sale through the box office system. Initially the purpose was, effectively, to build a mailing list - still a valuable use. But things have moved on.

Whether your system is a battered “Old Faithful” or the shiny “Latest Model”, under the bonnet there are nuggets of pure marketing gold.

About 15 years ago, in a meeting about segmentation, a marketing manager announced that he had “200,000 markets” - one for each customer. At the time we thought him mad. What a visionary. Here we are with dynamic emails, dynamic pricing and social media giving immediate, personalised access to customers. It is now genuinely possible to tailor the offer to the individual. What's usually missing is the time, the resources - and perhaps the courage - to do it.

Never fear: however unique we like to think we are, when it comes to being a consumer there are always a good number of people pretty similar to us - similar choices, similar attitudes to price, similar booking preferences. So even in an emerging world of personalisation, segmentation can make the biggest difference to your organisation since you bought a ticketing system.

I'm not suggesting we return to the mailing-list-only days. I'm suggesting we look under the bonnet of our ticketing data and find those nuggets. Box office data tells us far more than a name and address. Derived information - how often someone visits, how much they spend, how many they attend with, how far they travel, how far in advance they book - can all inform our marketing and earn the best possible return on budget.

Most segmentation in arts organisations revolves around genre: someone came to a ballet, so we send them more ballet. But if the goal is loyalty, genre-based segmentation isn't much use. There are two things to be aware of when we segment: propensity (how likely a customer is to come back) and interest. Genre is one way to match events to interest; in this piece we focus on propensity.

Building loyalty is a two-way street - the customer must get something back for their loyalty. So we need a segmentation that tells us where to invest the marketing budget: looking after those loyal to us, and not those who cost more than they generate.

Principally, there are three basic segments in every organisation's list:

• Your “best” customers
• Your “worst” customers
• Everything in between

So who are your best customers?

It depends entirely on what you're trying to achieve - a big question for arts organisations, where the primary motivation isn't always ticket revenue. For this piece we'll keep it to plain single-ticket sales (not subscribers or group buyers). So our “best” customers are those who come a lot; our worst are those who don't; and there's everything in between.

How do we define “a lot”?

Ask most arts marketers to define their best customers and the answer is usually “they come three or more times a year.” In reality that doesn't work. Which of these is the better customer?

• The customer who came three times a year, every year, for the last three years.
• The customer who attended nine times this year but had never been before.
• The customer who came once and brought twenty other people with them.

By the “three times a year” rule, only one of them counts as “best” - yet surely all three qualify. The way out is to stop having a fixed idea of a good customer, and instead define criteria to rank customers in order of importance. Once ranked, we can describe those at the top against those at the bottom.

Introducing Recency, Frequency & Value

Since data has been collected in computerised systems, three variables have been used: Recency (how many days since the last purchase), Frequency (how many visits in total) and Value (how much has been spent in total). Together - RFV - they rank customers, and they're almost impossible to beat for it. Other good systems exist, but this one uses data you already have, and every customer on your database can be categorised, identified and communicated with appropriately.

Categorising your customers

You may find easier or automated routes, but here's the manual version in a spreadsheet:

1. Export every customer with two columns: date of last visit, and total number of visits. (Not always easy to export - you may need the data whizz in your organisation.)
2. Sort by recency (date of last attendance), descending, and score each customer 1–5. With 100,000 customers, the most recent 20,000 score 5, the next 20,000 score 4, and so on.
3. Re-sort by total frequency and do the same: the 20,000 most frequent score 5, the least frequent score 1.

Now what?

Represent the customers in a matrix. Filter by the scores to see how many fall into each box.

You'll probably find few in Box 1 and many in Box 4, with Boxes 2 and 3 roughly even. Now we can describe them meaningfully:

And from a lot of ticketing analysis, their propensity to return:

Create your strategies

Simplistically, each segment gets a strategy. These differ by organisation (most marketers enjoy this bit), but as an example:

In simple terms, being able to identify and not market to Box 4 saves most organisations a great deal of money.

This is only the start of the journey - but once the foundation of a segmentation strategy is in place, there's a great deal you can build on it. In the next piece I look at the different behaviours each of these segments shows, and what that means for how, when and what we communicate to them.