When I am not working in the Big Smoke ( Sydney - the “big smoke” is a soubriquet for any large city, originally London, but in Australia it has something of a slang nuance - see for example Barry Popnik, and the straight facts at the inestimable Wikipedia), I live and work in a small historic country town a few hours drive West of the Smoke.

The town has a group of players which put on amateur theatricals once or twice a year, a much loved community event, probably very much the same as any small repertory group in any small town around the world. And the nearby larger town, nearby being about an hour’s drive - these are Australian distances you understand - also has a drama society.

We are enthusiastic supporters of these groups, good for the health of the community, good for us. But my professional instincts and interest also kick in, and I start to wonder what data analytics could contribute to these institutions and others that are so emblematic of small-country-town life.

And I have an abiding interest in the arts, and researching the arts..the application of Q methodology being a part of it (see, for example, Len Barchak’s work with Q for the Lake Charles Symphony Orchestra). But let’s set primary research aside for a time.

Like many arts groups they have a problem - low levels of patronage, high fixed costs (insurance etc), only a small marketing budget and no research funds.

So, with my data analytics hat on, what data do we have, how can we analyze it, what low-cost high-payoff strategies could we adopt to ensure a wider and continued patronage?

Well, from simple observation

• they are “preaching to the converted”
• the “core” is too small .. loyal, but too small
• there are “notable absences” .. those who you expect would/should be there, but are not .. people of standing in the community, people who represent particular sectors
• sporadic attempts have been made to attract a wider audience, but these have been diffuse and uncompelling
• it is probably reasonable to assume that awareness is low outside the core converts, but that there is general good will towards the enterprise
• repeat purchase has probably come about by accident .. those attending are not so much those “interested in theatre” but those who, for one reason or another, have attended in the past and repeat out of inertia or duty or, indeed, pleasurable expectations.
• perhaps 80% of theater-goers are repeat purchasers, not much new blood on any given occasion
• there is a strong “social circle” effect - stronger than just straight word-of-mouth. People invite others along to these events, they “make up a table” of their friends and relatives.
• mostly people are “satisfied” with the performances .. “satisficed” might be closer to it (to satisfice: to obtain an outcome that is good enough… In recent decades doubts have arisen about the view that in all rational decision-making the agent seeks the best result. Instead, it is argued, it is often rational to seek to satisfice i.e. to get a good result that is good enough although not necessarily the best. The term was introduced by Herbert A. Simon in his Models of Man 1957 ). 

  There does not appear to be a basis for believing that declining patronage is strongly linked to dissatisfaction, more to natural attrition (loyal theatregoers, as they drop off the twig,  not being replaced by new attendees)

There is enough (soft) information here to build a macro-flow model .. a model of individual and group decision making, for initial and repeat purchase.

So, without doing primary research, without building a macro-flow model, what is the best way to proceed given this explication of the status and mechanics, and  given a small budget?

“Structured shotgun” aka “Stratified Marketing”

“Structured shotgun” is the short answer. And hence the title:  “stratified marketing”.

We have some priors. Some uncertain beliefs about the likely payoffs of certain courses of action.  Is it optimal to choose the single “highest probable payoff” strategy? No, probably not.

There is an analogy here with portfolio (diversification) theory, and we must also consider the costs of being wrong. In this case we should probably allocate marketing effort across several somewhat uncorrelated strategies.

In practice this might mean

• a budget allocated to identifying and influencing disjoint opinion leaders - there need not be many, because of snowball effects - and putting effort into convincing them to become patrons, special table holders.. whatever.

  A stratified approach is taken within this - eg strata of farmers/agriculturalists, old money, in-town professionals, factory and infrastructure owners… This is a high-cost high-touch high-effort element of the overall strategy.

• a budget allocated to generalized awareness .. low cost (eg radio), mass media appeals : again, segmented by media (radio, leaflet drop, telephone calls..)
• efforts segmented by location .. in town, fringe, rural, niche and village communities
• success monitoring .. some exit interviews to determine source of business, and the effectiveness of the various strategies, intention to re-visit

None of this is rocket science. But we used a bit of structured common sense, examined what we thought we knew,  put a plan in place that reflected both our perceptions of likely payoffs and covered the field - so that next year, we would have more data and we could continue the virtuous circle.

Stratification - allocation of effort and budget across elements of the marketing plan, according to our perceived payoffs and lack of knowledge - is a very cost-efficient way of covering the bases, maximizing potential payoffs and simultaneously managing risk.

 I commend the idea to you.

And if you are in the fortunate position of having a large marketing budget, the principles still apply.

Stratification and experimentation can go hand in hand.. for example, it is entirely practicable to allocate advertising expenditures across media and across regions to learn about effects but simultaneously to work within an acceptable range of a presumed optimum.