The Gut Feel Economy
58% of companies make half or more of their regular business decisions on gut feel rather than data (BARC). The other 42% are mostly using data to confirm what the gut already decided. I'm being slightly unkind. Only slightly.
For startups the number is almost certainly higher. When you're moving fast, "I think" is shorter than "let me check," and the calendar wins.
Meanwhile, the data on data-driven decisions is unflatteringly clear:
- Data-driven organisations show 25% higher EBITDA (McKinsey)
- They are 3x more likely to report significant decision-making improvements (PwC)
- Companies that leverage data in decision-making are 23x more likely to acquire customers and 6x more likely to retain them (McKinsey)
25% higher EBITDA isn't a margin. That's a different company.
The Perishable Insight Problem
Here's the standard plot:
- Founder has 10 great customer conversations this week
- Notes are scattered across WhatsApp, email, a notebook, and the inside of one head
- Friday team meeting — founder shares 3 insights from memory
- Product decision made on the version of events that survived the week
- Six months later: "no market need"
But the market need was there. It just didn't survive the trip from the conversation to the decision.
Customer insights are perishable. Like fish, they keep for about 48 hours. After that, they aren't insights anymore — they're guesses wearing the confidence of insights.
The Data You're Not Capturing
Most startups aren't even collecting what they would need to make good decisions:
- Customer conversations — notes taken sporadically, key details lost within days
- Sales pipeline reality — forecasts built on rep optimism, not buyer behaviour
- Churn signals — 70% of churning customers showed warning signs 2 weeks before leaving, but nobody was watching the activity patterns
- Feature usage — building what you think users want vs. what they actually open
- Revenue leakage — missed renewals, untracked discounts, invoices sent late and remembered later
The data exists. It's flowing through your email, your calls, your product usage logs. It just isn't arriving anywhere useful. Imagine a kitchen where every ingredient is delivered to a different room and nobody's allowed to walk between them. That's most startup ops.
Why Small Teams Have an Advantage
Counterintuitively, small teams should be better at this than large ones. With 3-5 people:
- Faster feedback loops — insight to action in hours, not quarters
- Fewer handoffs — the person who talks to customers also builds the product
- Lower noise — fewer competing versions of what the customer "really meant"
- Easier systems — you don't need a data warehouse, you need a CRM that doesn't require a manual
The problem is that small teams are so busy executing that they never build the systems to capture what they're already learning. The advantage is sitting on the desk. Nobody has time to pick it up.
How to Start
You don't need a data team. You need three things:
1. Auto-capture everything
Stop relying on humans to log data. Humans are excellent at many things; remembering to update fields after a 6pm sales call is not one of them. Use tools that capture automatically:
- Email and meeting content → summarised and tagged
- Sales interactions → pipeline updated without manual entry
- Customer communications → unified timeline across channels
2. Build decision triggers, not dashboards
Dashboards are where data goes to die in peace — beautiful, untouched, reviewed once a quarter by people who are mostly pretending. Build alerts instead:
- Customer activity drops 60%? → Notification
- Lead uncontacted for 24 hours? → Escalation
- Invoice unpaid for 30 days? → Automated follow-up
The dashboard waits for you. The alert finds you. Only one of those is going to save your week.
3. Review weekly, not quarterly
A weekly 30-minute review of actual data — pipeline, customer activity, revenue — catches problems 12 weeks earlier than a quarterly review. In a startup, 12 weeks is the difference between course-correcting and writing a "lessons learned" post on LinkedIn.
The Compounding Effect
Good data compounds. Every customer conversation captured makes the next product decision better. Every sales interaction logged makes the forecast a little less fictional. Every pattern recognised makes the system marginally less surprised by Mondays.
Bad data compounds too. Every missed note, every unlogged call, every shrug-shaped decision adds uncertainty. Over 12 months, the gap between the data-driven startup and the gut-feel one stops being a gap you can bridge. It becomes a gap you watch from across.
25% higher EBITDA. That isn't a rounding error. That's survival, dressed up as a percentage.
Where My Co-Founder Stops Me
I showed a version of this to Jabez, our CEO, expecting a nod. I got a counterargument.
He doesn't think the data is wrong. He thinks I'm about to use it as an excuse. Our CEO has a line on this that I think keeps the team honest: "A good decision today beats a perfect one next quarter."
His case is uncomfortably good. Waiting for the fully-evidenced, no-regrets call is often just procrastination wearing a nicer suit — the same instinct as the gut-feel founder, only dressed in a lab coat. Twenty fast, roughly-right decisions will lap one perfect one, because volume is where the learning actually lives. The team that runs one careful experiment a quarter loses to the team that runs twenty scrappy ones. Especially the half that go wrong — those are the ones that teach you something.
And here's the part I keep turning over. A decision made on Monday's data and shipped is often worth more than a better decision made on Thursday, because by Thursday the data has quietly moved underneath you. Insights aren't the only thing that's perishable. So are decisions.
So Which One Wins
Neither, and I've stopped pretending that's a cop-out.
Data without speed is a beautifully sourced conclusion that arrives after the meeting has ended. Speed without data is confidence with the safety off. The founders who pull ahead aren't the most careful or the most decisive — they just quietly arranged their operation so that being fast and being informed stopped being a trade-off.
That's the part nobody frames as exciting: it's a systems problem, not a personality one. When the customer conversations, the pipeline, and the revenue capture themselves, "let me check" and "let me decide" stop being different speeds. The data is already on the desk when the moment arrives. You get to move at Jabez's pace with my paperwork — which, as far as I can tell, is the only version of this argument where nobody loses.
25% higher EBITDA, or twenty shots instead of one. It turns out you were never supposed to choose. You were supposed to build the desk where both are already sitting.
RUBL auto-captures every customer interaction and surfaces the insights you need. No manual data entry. See how it works.