The Customer Base Is Not the Community

Suppose 600 people buy tickets through the Duncanville Arts Foundation’s Box Office.

We can learn a great deal from them.

What they bought.

What they paid.

Whether they attended.

Whether they purchased again.

Those are valuable observations.

But those 600 customers do not automatically tell us what everyone in Duncanville wants.

That distinction may be one of the most important safeguards in the Cultural Interest Study.

The Box Office can be accurate and still not be representative

There are two different questions here.

First:

Did we accurately record the transactions that happened through the Box Office?

Second:

Do those customers represent the entire population of Duncanville?

Those are not the same question.

The Box Office may give us a complete record of the Foundation’s ticket transactions during a particular period.

Every ticket could be recorded correctly.

Every price could be accurate.

Every customer could be accounted for.

We still cannot assume those customers behave like every resident of the city.

Buyers select themselves

People do not enter the Box Office because the study randomly selected them from every household in Duncanville.

They enter because they encountered an offer and decided to act.

That means our customer base is shaped by the conditions surrounding the offer.

Change those conditions and we may observe different customers.

We are observing the people who encountered our offers and chose to respond to them.

That is useful evidence.

It is not the same thing as randomly sampling the city.

Representative samples require a different design

Survey researchers make this distinction carefully.

Pew Research Center explains that when people enter a study through nonrandom or opt-in methods, researchers do not know the probability that every member of the larger population had of being included.

If the people who participate differ from the broader population in important ways, their responses may not represent the whole population.

Compare that with the National Endowment for the Arts’ Survey of Public Participation in the Arts.

The 2022 survey was conducted as a supplement to the U.S. Census Bureau’s Current Population Survey.

An adult was randomly selected from sampled households, and the study used a formal sampling and weighting process to estimate arts participation among U.S. adults.

That is a different research design from a customer database.

Our pre-validation survey is another kind of evidence

The Cultural Interest Study begins before the transaction.

Pre-validation allows us to ask people what interests them before we ask them to buy.

That is important because it lets us compare:

But we should apply the same discipline to the survey.

A survey does not become representative simply because we call it a survey.

Its ability to describe all Duncanville residents depends on how participants were recruited, who had an opportunity to respond, who actually responded, and how the results were analyzed.

Every dataset should be allowed to say only what its design supports.

The comparison is still powerful

Suppose a pre-validation survey finds substantial expressed interest in theater.

Then the Foundation offers a theater experience.

Purchases are weak.

That difference is valuable.

But it does not automatically mean the survey was wrong.

Perhaps:

The gap between expressed interest and actual purchasing becomes another question to investigate.

Our programming shapes what we can observe

There is another limitation that is easy to overlook.

We can only observe purchases for things we actually offer.

Suppose the Foundation produces:

And no dance performances.

At the end of the year, the Box Office may contain many music transactions and no dance transactions.

We cannot conclude that there is no demand for dance.

We never offered anyone the opportunity to buy it.

What we choose to test determines part of what the study is capable of discovering.

Absence of a transaction needs an opportunity

This gives us an important research rule.

If nobody buys opera tickets when we never offer opera, we learned nothing about opera demand.

If we offer opera once, with little promotion, on an inconvenient night, and sell poorly, we learned something about that test.

We still may not know the size of the larger opera market.

If we test several offers under different conditions and repeatedly observe weak purchasing, our confidence can increase.

The size of the conclusion should grow with the evidence.

Marketing also shapes the sample

Suppose an event is promoted primarily through the Foundation’s existing email list.

Many buyers may already know the organization.

Now suppose another event is promoted through a school district, neighborhood businesses, and regional advertising.

The people who encounter those offers may be different.

That means marketing does more than sell tickets.

During a study like this, marketing also affects who has an opportunity to enter the observed customer pool.

Who sees the offer can influence who appears in the data.

Location can shape the observed audience

The same is true of place.

An event at a restaurant may attract one group of people.

A senior center may attract another.

A park may produce another pattern.

A temporary gallery may reach people who would not attend the same program somewhere else.

That is one reason the Cultural Interest Study benefits from testing activity in different places.

We are not only testing the art.

We may also be testing the relationship between an arts experience and its setting.

Price can shape who enters the customer base

A $5 ticket and a $75 ticket may not produce the same set of buyers.

That does not mean one audience is more important.

It means the price is one of the conditions shaping the observed transaction.

If we later describe the people who purchased, we should remember how they arrived in the dataset.

They responded to a particular product at a particular price under particular conditions.

Do not turn customer demographics into city demographics

Suppose the study eventually collects demographic information from customers.

Perhaps we discover that a particular age group accounts for a large share of observed purchases.

That may be useful for understanding our customers.

It does not mean that age group represents the same share of Duncanville’s overall interest in the arts.

The difference matters.

The first statement describes our evidence.

The second makes a population claim the transactions alone cannot support.

Words can make a claim too large

Reporting language matters.

Consider:

Or:

Or:

The claim can grow as the evidence grows

Caution does not mean the study can never reach strong conclusions.

It means those conclusions have to be earned.

One test may support a narrow observation.

Several similar tests may reveal a pattern.

Different audiences, locations, prices, and dates may strengthen or weaken that pattern.

Repeated purchasing may give us evidence of persistent demand.

Broader survey work may help us understand whether the pattern extends beyond current customers.

Stronger evidence allows stronger language.

The people who did not buy also matter

A customer database contains customers.

By definition, it tells us less about people who never entered the transaction.

Some may have seen the offer and rejected it.

Some may have been interested but unable to attend.

Some may never have heard about it.

Some may spend substantially on the arts elsewhere.

Some may have little interest at all.

They are part of the community, too.

The Box Office cannot describe them simply because they are absent from it.

Outside spending complicates the picture

Duncanville residents can spend money on the arts without spending it with the Duncanville Arts Foundation.

They may buy:

Those transactions are part of residents’ arts buying behavior.

They may never appear in our Box Office.

That is one reason the Cultural Interest Study needs more than Foundation ticket sales if its larger question is where Duncanville residents spend money on arts-related activity.

The Box Office is still one of our strongest tools

None of this makes transaction data less valuable.

Quite the opposite.

A purchase is strong behavioral evidence.

It tells us someone encountered an actual offer and accepted an actual price.

Repeat purchases give us stronger evidence over time.

Attendance gives us another behavior.

Spending around the experience gives us another.

We simply need to keep the evidence attached to the people and conditions that produced it.

The strength of transaction data comes from what actually happened, not from pretending the transactions represent people we did not observe.

Different sources answer different questions

The Cultural Interest Study becomes stronger when we stop asking every dataset to answer every question.

Each source adds another piece.

None should casually be substituted for another.

We should define who each finding describes

This gives us a useful reporting habit.

Before publishing a finding, ask:

That simple discipline can prevent a small observation from turning into an oversized conclusion.

The study does not need one answer for everybody

The goal is not to discover one universal Duncanville preference.

There may not be one.

We may instead discover several observable markets.

Different buyers.

Different art forms.

Different prices.

Different places.

Different frequencies of purchase.

Different reasons for participating.

That is a more realistic picture of an arts economy.

The reporting rule

By the time the Cultural Interest Study produces formal findings, every major claim should answer one additional question:

Who does this evidence actually represent?

If the answer is purchasers in one test, say purchasers in one test.

If it is repeat customers across twelve months, say that.

If a carefully designed survey supports a broader estimate, explain the survey and its limitations.

If we do not know, say that, too.

The language should never outrun the evidence.

Subscribe to Ron Thompson’s Field Notes

Occasional notes on art, culture, creative economies, and the work of building cultural capacity from the ground up.

Subscribe to Field Notes

By sending the subscription request, you agree to receive Ron Thompson’s Field Notes by email. Your email address will be used to manage and deliver your subscription and will not be sold. You may unsubscribe at any time by replying “unsubscribe.”