Interest will not make payroll

In late 2019, all the walls and windows of our King St office were covered in post-it notes. Not a few. Hundreds. Nearly every pane and wall, colour-coded by user journeys and product capability. If you'd walked in cold, you'd have assumed we were planning a coup.

We were building the beta of Facilitate, and we were doing it by the book, following tech startup best practice all the way:

• Created assumptions then tested them in customer interviews
• Created our business model canvas (and even the lean model canvas variant)
• Built a Wizard of Oz version to show what the product could become
• Designed the real thing alongside the people who would use it: a dozen pilot customers, learning and development consultants, government departments and Australian enterprises (some of them household names). Each one had agreed to pay for the pilot. Not much, but something. We wanted the small friction of a transaction, an agreement, a signature, because that's what the books said separated a real customer from a polite one.

Then we built what they'd asked for. Four things:

1. Multi-person immersive learning environments
2. A scenario practice tool where a learner could rehearse a hard conversation with a peer or a coach, including ten unique 3D environments for that bad boy: offices, hospitals, a public-speaking stage, industrial sites
3. A Q&A capability
4. A 360 media viewer

The design took weeks and weeks. The build took months and months. A small team of designers, developers, an architect, an organisational psychologist, a project manager, working day in and day out on a plan we'd all agreed to and that our customers had told us they wanted.

We executed it cleanly. The team did a phenomenal job. I'm still proud of that.

---

Here is what the plan did to me. It made me certain.

Twelve paying pilots is a number you can put in a deck. Well-known logos are a thing you can say to a board and investors. A product designed with your customers (your 'friendlies') is a sentence that sounds like product-market fit even when it isn't. I wasn't exaggerating any of it. I believed it, and the belief carried everything. It held up the forecast, the hiring, the hours, and my own sense that the hard years were about to pay off.

So when the pilot went live and the usage data came in, I kept reading it the way I'd been reading the post-it notes.

The scenario practice tool, the thing we'd built ten environments for, the capability that had been held up as a critical value-adding differentiator by customers: nobody used it. Not a little. Not at all. Meanwhile the 360 viewer, the simplest thing in the whole build, was getting usage. And the customers using it weren't asking for more viewer. They were asking for interactivity inside the scenes, which we hadn't built and which would take another six months.

Then COVID arrived. We were selling software that required a headset you put on your face. All of a sudden, a device that was seen as 'the next computing platform' became as appealing as licking a toilet seat in an airport restroom. In a fortnight every pilot customer, every single one, told us this was no longer on their agenda. That wasn't a judgement on the product. It was a reflection of the uncertainty and fear that engulfed everything at that time. VR was instantly put in the 'save for another day' list of innovations to consider.

Months of no revenue. A team that had done everything right, looking at me for a direction. A playbook we'd followed line by line, and a gut punch at the end of it. It was a year before we got momentum back, and we got it back by building a desktop version none of the original pilot customers had asked for, because in the depths of lockdown we had no idea when anyone would wear a headset again.

Eventually one of the twelve pilots subscribed. One out of twelve. That's the number I'd put in the deck now.

---

If I could sit with the man in that office, surrounded by post-it notes, I'd tell him three things.

The nominal fee was the wrong signal. We asked for something to prove commitment, and we got exactly what we asked for: something. A company that will pay a small amount to try your product is telling you the problem is interesting. A company that will pay what the problem actually costs them is telling you it's urgent. We measured interest and called it demand. If a customer won't pay a rate that reflects the pain, the pain isn't big enough, and the kindest thing they can do is say so, and the kindest thing you can do is move on.

One capability, not four. Every feature we added to the pilot diluted the one question that mattered: is this problem burning? We took the macro problems surrounding training (training isn't standardised, it's disengaging, it doesn't scale) and attempted to solve them with an immersive solution. Four features gave twelve customers twelve different reasons to say yes in a workshop, and none of them had to mean it. One feature focused on solving one specific burning problem would have forced the real conversation months earlier.

And some things are out of your control. COVID took out a market we'd spent three years building toward, and no interview, no pilot, no amount of post-it notes would have seen it coming. You can do everything right and still get clobbered.

The founders whose books I'd read had friendlies too. Theirs converted. That's the whole difference, and it's why their advice reads so clean. Mine didn't, and the lesson I got instead is the one I'd want in your hands before the post-it notes go up.

---

One thing to try this fortnight. Pick the warmest conversation in your pipeline. Before the next demo, before the next feature, ask them two questions and don't rescue them from the silence:

1. What does this problem cost you, in money or hours or sleep, right now?
2. How have you tried solving this in the past or are actively trying to solve it now?

If they can't put a number on it, or the number is small: you have interest. If they haven't tried to solve it in the past or don't have active budget to do so now: you have interest. Interest feels good. It is not a purchase, and it will not make payroll.