The Next Move Belongs to the Market

Yesterday I was sitting in my office with football on. Not unusual. What was unusual was not being able to think of anything I actually wanted to build.

I don't mean nothing I could build. That's never the problem. I can always find something. Another dashboard. Another research pass. Another workflow. Another field we could enrich. Another screen that would make something 7% prettier. Another system to make the other system more systematic.

I have developed a fairly impressive ability to turn almost any uncomfortable business question into a software project.

Far enough

But yesterday I looked around at everything I've been working on lately and had a strange realization: a lot of it was far enough. Not finished. Not perfect. Far enough.

And I hated it.

It's been an interesting week. I've got a lot of things moving at once. A local advertising concept connecting businesses with restaurants, breweries and other venues. A shipping-conversion opportunity where we're identifying companies that may be good candidates to move parcel volume. Different versions of territory intelligence for sales teams. A couple of companies interested in seeing what we're building. A potential in-person implementation workshop where the entire point is that the team leaves with actual work running, not a binder full of "AI strategy."

Different markets. Different products. Different buyers. But I noticed the same thing happening across almost all of them. We had reached the point where the next important piece of information couldn't come from Elevare.

That changes everything.

The answers live outside the building

Take the local advertising business. We've already done a ridiculous amount of market work. Thousands of venues. Thousands of potential advertisers. Tens of thousands of possible matches between them. Pricing research. Print vendors. Different operating models. Different packages.

We can absolutely keep going. We could identify another 5,000 advertisers. Score every dentist in central Indiana. Build a gorgeous marketplace. Add automated matching. Create a perfect venue portal. Hell, give me a slow weekend and I'll probably have an advertiser-to-coffee-shop compatibility algorithm by Sunday night.

But none of those answer the most important questions anymore. Will a local business actually pay for this? Will the venue say yes? Does the pricing feel right? Can we get the first board sold?

Those answers don't live in another dataset. They live outside the building.

Same thing with the shipping project. We built a universe of more than 70,000 companies that plausibly ship meaningful parcel volume. Then we narrowed. Then we looked for stronger evidence. Then we started examining websites, shipping policies, operating clues and everything else we could reasonably find. Useful work.

But here's the trap. Going from 70,000 possibilities to 500 incredibly researched possibilities feels like progress. Going from 500 to 100 perfectly researched possibilities feels like even more progress. But the business isn't created when we identify the 100th perfect account. The business is created when one of them converts.

There is a point where the next enrichment field doesn't make the opportunity any more real. A conversation does.

Is the next unknown internal or external?

I'm starting to think this is one of the most important distinctions in building anything.

If the next unknown is internal, keep working. Can we technically do this? Do the economics work? Can we identify the customer? Can we build a usable version? Do we understand the problem? Do we have enough evidence to make a reasonable offer? Those are things you can investigate. Build. Analyze. Model. Test internally.

But eventually the questions change. Will somebody use it? What will they pay for it? Does this actually solve the thing they care about? Will the salesperson trust the recommendation? Will the customer change what they're already doing?

And once those become the important questions, another week inside the workshop isn't discipline. It's avoidance.

Building feels safe

I know this because I'm really good at it. Building feels safe. There is always something to do. You can work twelve hours and point to what changed. Look: more records, more features, more research, better UI, cleaner architecture, more documentation. Another green checkmark.

The market is much less accommodating. You can spend twelve hours talking to people and discover the thing you spent three weeks building is not particularly important to them.

How rude. Also: uncomfortably, extremely useful.

The ten-minute reaction

I've watched myself do this with sales intelligence too. Say I build a system for someone who manages a territory. The first version might tell them: here are all the accounts. Then: here are the better accounts. Then: here are the accounts near where you're already going. Then: here's what changed at those accounts. Then: here's who you might want to talk to.

I can keep making that smarter indefinitely. But eventually I need the person who actually drives the territory to open it on Monday morning and tell me, "Yeah, this is useful." Or, "No. I don't care about that. What I actually need is this."

That ten-minute reaction can be worth more than another week of development. Because now the next thing I build is anchored in reality.

Implementation creates feedback

I've been thinking about this with consulting too. There's a reason I'm increasingly drawn to implementation rather than advice.

Imagine selling a company a day of training. You could spend six hours teaching them everything there is to know about the latest tools. Very impressive. Everyone takes notes. People nod. Someone gets a nice lunch. Monday arrives and nothing changes.

Or you can structure the day around one requirement: everybody leaves with something working. A repeated process is automated. A project has the right context. Company information is connected. A tedious administrative job is actually easier tomorrow morning than it was yesterday.

Now we have a result. And more importantly, reality starts giving us information. Did they use it? Did it save time? Where did it break? What did we misunderstand? What should we improve?

Implementation creates feedback. Advice often creates another meeting. And I hate meetings. (Proudly team camera-off.)

Building is not a substitute for reality

That's probably the bigger lesson I've been learning lately. Building is useful because it gets you to reality. Building is not a substitute for reality. Those sound almost identical. They are not.

You need enough product to test the idea. Enough research to make the call intelligently. Enough intelligence to prioritize the right account. Enough infrastructure that the thing doesn't fall over immediately.

But "enough" is doing a lot of work in those sentences. And builders are terrible at enough. We like finished. We like polished. We like comprehensive. We like clever. The customer may just want something that saves them 45 minutes tomorrow.

The chain has to end somewhere

This also changes how I think about company intelligence. Last week I wrote about how you can drown in information. That problem is real. But there is another failure mode on the other side of it.

You can take all that information, analyze it, filter it, turn it into real intelligence… and then still refuse to act. At that point you've built an extremely sophisticated procrastination machine. The inertia of the speculation station is real.

The chain only works if it ends somewhere:

Information → Analysis → Intelligence → Action → Response

That last step is the one that gets forgotten. The customer responds. The salesperson reacts. The market moves. Something works. Something doesn't. And that response becomes new information.

That's the loop. Not build, build, build, build, launch. It's:

Build enough → act → learn → build what the learning justifies

Much faster. Much cheaper. Probably much less emotionally comfortable.

Cheap building removed a safeguard

I think this matters especially now because the cost of building has collapsed. It used to be expensive enough that people had to be selective. Software took months. Research took teams. Prototypes took real money. You naturally ran into constraints.

Now a motivated person can produce an absurd amount of stuff in a week. That's an incredible advantage. It also removes one of the old safeguards. Nobody is standing there saying, "Josh, are you sure we should spend $80,000 developing this feature?"

Instead it's, "Sure, we can probably knock that out tonight." (Narrator: it did not take one night.)

One-shotting it in a night sounds wonderful. Until you realize cheap things can still waste time. And time is still expensive.

The rule

So I'm trying to build a new rule into the way I work:

When the next unknown is external, stop doing internal work.

Send the text. Book the demo. Show the ugly version. Ask for the money. Put it in someone's hands. Let somebody tell you it's wrong. Then go back to work.

There will be plenty to build after that. There always is.

Who owns the next move

I still don't particularly enjoy the part where things are far enough and I have to wait for another human being to react. I spent part of yesterday sitting in my office apparently waiting for the universe to assign me another task. It did not. Football remained deeply unconcerned with my pipeline.

But maybe that discomfort is useful. Maybe it's the feeling of reaching the boundary between building something and actually building a business. Because eventually the next move isn't yours. It's the customer's. And the smartest thing you can build at that point might be nothing at all.

Know what you need to learn next. If the answer lives inside the company, go find it. If the answer lives with the customer, get out of the building. An ironic discovery for someone with a career in outside sales.

The difference between endless building and real progress may be knowing who owns the next move.

This is why our workflows and AI agents work is built around one standard: the team leaves with something running, not a binder. New to the series? Start with What Is an Intelligence Estate.

Questions people ask about this