Every small business owner has a dozen ideas a week. Some come from customers. Some come from market signals. Some come from 2am thinking about why last quarter was weaker than you expected. Most of them are fine, some might work, and a couple might be genuinely important. The problem is knowing which is which when you're also running the business.
We built something to help with that. Not because we needed AI to make decisions for us, but because we needed something that would reflect our actual priorities back to us in a systematic way, and then push back honestly when we were drifting.
How we framed the problem
The first step was defining what "good" actually means for us. That sounds obvious until you try to do it. If you ask ten people in a business what the priority is, you get seven different answers, and three people who think it's a trick question.
What we wanted was something that could score a new idea against these weighted criteria: margins and unit economics (the thing that keeps the business alive), brand fit and positioning (the thing that keeps us us), team capacity and risk (the thing that doesn't burn people out), and market signal strength (whether anyone actually wants this).
We weighted those criteria differently depending on the time of year and the state of the business. When cash is tight, unit economics gets a heavier weight. When we're stable and can afford to build something interesting, brand fit and positioning gets more leverage. That weighting isn't random. It's us having said: these are our actual priorities right now.
The tool takes a proposal, measures it against those weighted criteria, and gives it a score. More importantly, it explains the reasoning behind the score, which is where the real value lives. Because the score isn't the point. The thinking that produced the score is.
What happened when we actually used it
The first few weeks were interesting. We threw half-formed ideas at it, expecting it to rubber-stamp them or kill them. What actually happened was slower and more useful. The tool would score an idea as worth exploring, but the reasoning would reveal a gap we hadn't articulated. "This could work on margins, but it assumes we maintain current pricing. If we drop price by 10 per cent to compete, the math falls apart." That's not the tool deciding the idea is bad. It's the tool forcing us to be specific about the assumptions we're hiding inside our enthusiasm.
And here's the part that surprised me most: when the tool said an idea was actually strong, that carried weight. Because it wasn't coming from optimism or pattern-matching or the fact that someone made a good pitch. It was coming from the system we built together to represent our actual values. That meant "the tool says this is worth doing" felt different than "I think this is worth doing."
The other surprise was the pushback. We built in the capacity to explain when a score didn't feel right, which means we're constantly refining the weighting and the criteria. That's the maintenance piece. A strategic tool that never changes has already stopped being useful.
Why this is different from just having a smarter owner
According to McKinsey's 2025 State of AI survey, 64% of organisations report that AI is driving real revenue. But the gap between adoption and meaningful deployment remains wide, especially for small businesses. A smarter owner probably could run these calculations in their head, if they had enough mental bandwidth and weren't tired. The value of the tool isn't that it thinks better than a human. It's that it's consistent, it's explicit about assumptions, and it won't skip the boring criteria because it's excited about a shiny idea.
Humans are pattern-matching machines. We're brilliant at seeing connections and possibilities. We're also terrible at remembering to apply criteria we said mattered when we're emotionally invested in an idea. The tool doesn't have that problem. It applies the criteria every single time, even when the idea is your own or it's politically important or it came from someone senior.
That impartiality is most valuable when you're tired, when the business is under pressure, when you're tempted to try something that violates your own strategy because it might be the quick fix you need right now.
How this scales for a small team
This started as a thought exercise with a spreadsheet. Then it became a Claude conversation where we'd paste in proposals and it would score them. Then we turned it into a basic interface with sliders for the weightings so you could see how different priorities would change the scores. That interface lived inside a low-code tool and sat next to our main business data.
Forbes reported that by the end of 2026, more than 80% of small businesses will be using AI-powered marketing tools. The tools exist. The question is whether they're being used strategically or just bolted on. The point is: you don't need anything sophisticated to build this. You need to articulate your criteria, define the weighting, and have something that scores consistently. The medium doesn't matter. What matters is that it exists outside your head.
As the team grows, this becomes more important, not less. Because when you're six people, you can hold the strategy in the relationships and the conversations. When you're twelve or thirty, you need the strategy written down somewhere that's not someone's email. This tool is the start of that.
The meta-benefit: what building this taught us
The most valuable thing we got wasn't the scores. It was the clarity. Building the tool forced us to say what we actually care about, in order of importance, with numbers attached. That's the work that usually doesn't happen. Strategies stay vague, criteria stay implicit, and decisions stay arbitrary.
Now when a new idea comes in, or when someone asks "should we do X," we have something to point to that's not "I have a feeling about this." We have "here's how it scores against our actual priorities, and here's why." That's a completely different conversation.
| Criterion | What we measure | Why it matters | Current weight |
|---|---|---|---|
| Unit economics | Contribution margin, payback period, cash impact | The business has to survive | 30% |
| Brand fit | Consistency with positioning, brand language, customer expectations | We don't want to be everything to everyone | 25% |
| Team capacity | Time required, skill fit, burnout risk, contingency planning | People are the constraint | 25% |
| Market signal | Customer requests, channel interest, competitive pressure, category trends | We follow signals, not hunches | 20% |
The weightings shift quarterly based on our business state. This table is this quarter's version. It'll be different when cash is tighter or when we're scaling.
Frequently asked questions
Does this tool make decisions for you?
No. It scores decisions systematically and explains the reasoning. We still decide. But now we decide with the same criteria applied every time, which means we decide better.
What happens when you disagree with the score?
You should disagree. If you consistently disagree, that means either the criteria are wrong or the weighting is wrong, and you need to fix it. The tool is a reflection of your strategy, not a replacement for it.
How often do the criteria change?
We review every quarter. The framework stays stable, but the weighting shifts depending on what the business needs most right now. When you're profitable and want to grow, the priorities are different than when you're in cash-squeeze mode.
Can this work for other drinks businesses?
The structure works for any business. The criteria and weighting would need to reflect your actual priorities. The value is in the clarity and consistency, not in copying our numbers.