By Mark Parsons, Director Events Intelligence
At the recent Indy Summit, I moderated a roundtable on the trade-offs between agility and scale for independent organisers. The discussion brought out both the freedom to experiment and the cost of that freedom: more options, more complexity and the psychological burden of things going wrong more often. Several organisers shared examples of initiatives they felt simply wouldn’t have been possible under alternative ownership. Their agility had helped them find new markets and improve the fit between their products and the communities they served.
The discussion also left me thinking about what survives all that experimentation. When we try something new, does the business become better informed, or simply busier? An independent organiser can have an idea on Tuesday and test it by Friday, which matters when a market is changing or an established event is overlooking part of its community. But in my experience, a business that acts quickly without examining the result can also repeat its mistakes remarkably efficiently.
Designing the experiment
I think this puts data at the heart of how we design an experiment. Before we try something, we should know which decision we’re trying to make, what we’re uncertain about and what evidence would change our view. That gives us a learning loop: a question, a test, a response and a better-informed next attempt. The lasting value comes from building evidence that improves our judgement, with each experiment giving the next one a better starting point.
Consider a hypothetical organiser hearing that procurement directors want more useful conversations with their peers. That suggests a customer need, but the commercial question is whether enough directors will pay enough, often enough, to support a recurring product. A small pilot could begin by testing whether they’ll give the proposition their time. Before sending invitations, the team agrees what response would justify a paid follow-up and how much time and money it can afford to spend finding out.
Reading the results
Suppose the team invites 80 relevant people to a free pilot, receives 20 confirmations and attracts 14 participants. That tells us something about the appeal of this particular offer to this particular group. If nine subsequently pay the intended price for another session, we have stronger evidence of demand. Whether that demand supports a viable product still depends on recruitment effort, delivery costs and staff time. If filling the room relies on the founder spending two days persuading people they know personally, that belongs in the assessment too.
The data needs to connect the offer, the audience, the response and the cost. We should retain the full group approached, including those who didn’t book, so that 14 people attended has a meaningful denominator. Recording the invitation, price, format and timing gives us the context needed to interpret their behaviour. Conversations with participants and people who declined can help explain what the numbers leave uncertain. These are small numbers, though, so our conclusions should remain provisional.
The next experiment should address whichever uncertainty matters most to the commercial decision. That might mean approaching a less familiar audience, testing a different price or changing the format to reduce delivery costs. Keeping other conditions reasonably consistent, where practical, makes the result easier to interpret. A disappointing response needs the same care: it could reflect a weak proposition, an unclear invitation or poor timing. Failure earns its keep when the evidence helps us decide what to change or where to stop spending.
Testing within the annual cycle
Truthfully, I think the annual show cadence makes many organisers understandably cautious. When a major opportunity comes round once a year, there may be limited room to recover from a poor decision. Yet smaller tests can sit within the marketing outreach and customer engagement we already do, allowing us to investigate an idea before committing substantial effort. Our roundtable made clear that many independent organisers are already doing this. The discipline is to give each test a clear question and a limit on the resources we’ll spend answering it, so experimentation itself doesn’t become another source of overload.
Making this practical requires modest additions to existing systems. I’d start with one experiment, one owner, a decision date and a short account of what we expect to learn. Use existing CRM tags to connect the approach with customer responses and commercial outcomes, keeping the relevant invitation, costs and feedback alongside them. At the review, compare what happened with what we expected and record the decision to continue, change or stop. Include the reasoning and its limitations while they’re still fresh enough for someone else to understand.
Turning experiments into data assets
This is where experiments can become data assets. A database records who came; a connected history of offers, responses, costs and decisions helps the business judge what to do next. Its value depends on colleagues being able to find the evidence, understand its context and apply it to a relevant decision. If the useful knowledge stays in the head of the person who ran the experiment, much of that value remains unavailable to the business.
Scale changes the economics of this knowledge because the cost of discovering an approach can earn a return more than once. A useful process may be repeated across editions, adapted for adjacent communities or taught to colleagues who weren’t involved in the original experiment. But the evidence travels with conditions: a format that works for established customers in one market may need another test elsewhere. Keeping those conditions visible helps the next team judge what it can reasonably reuse. Agility and scale reinforce each other when we can transfer both an approach and the evidence needed to adapt it.
When the learning loop disappears
Growth also creates a risk that the process survives while the learning loop disappears. Teams inherit the format, the original assumptions fade and “this is how we do it” becomes sufficient explanation. Recording which customer need an approach served, and what remains uncertain, gives colleagues a basis for questioning it as the market changes. The knowledge worth scaling includes the ability to recognise when a familiar approach needs another look.
The practical test is whether the record helps the next person make a better decision. If capturing it takes more effort than running the experiment, simplify what you collect. An organiser can begin with one complete learning loop and build from there, using each attempt to improve both its evidence and its judgement. What decision will your next experiment help you make, and what data will you need to make it?
