Many small businesses are struggling with Conversion Rate Optimisation (CRO), despite knowing its importance. Alon Abraham, co-founder and Head of Strategy at digital growth consultancy TAG, highlights that most small businesses are not doing CRO correctly.
They’ve read the blogs, moved a button, swapped a headline, and tried a new hero image, but nothing changes. Revenue remains the same, and the site doesn’t perform better.
The issue isn’t the tactic, but the approach. CRO only works when it’s a program, not a project. Businesses doing it well make small changes, more often, with a structured way of deciding what works and what doesn’t.
A one-off tweak can’t tell you anything. You change the headline, sales go up, and you declare victory, but you have no idea if it was the headline or something else.
Real CRO is a loop: hypothesis, test, measure, decide, and repeat. Done well, you know with statistical confidence that something worked, and more importantly, you know why. That learning compounds.
There are three reasons small business CRO fails. Firstly, most small-business owners obsess over the homepage, which is rarely where conversions are won or lost.
For e-commerce, it’s the product page, the cart, and the checkout. For services, it’s the enquiry form and the pricing page. Start where the money actually gets made.
Secondly, you’re running tests you can’t learn from. If you’ve got 500 visitors a month and you run an A/B test for two weeks, you’ll get a result, but that result will be noise.
You need enough volume to see a real signal, and if you don’t have it, testing isn’t your path – qualitative work is. Watch session recordings, run user interviews, and read your support tickets.
The insights are already there if you look. This is where many small businesses go wrong, as they tend to ‘do CRO’ for a month and then stop.
But the entire point is the compounding. A three per cent lift here, a five per cent lift there, a two per cent lift somewhere else – individually they’re boring, but stacked across a year, they’re transformational.
Related: Shared offices now offer extra services
Start with a question, not a tactic. “Why are people abandoning the cart?” is a better starting point than “let’s test a red button”.
Your analytics, customer service inbox, and session recordings will tell you where the real leaks are. Write down what you think is happening before you change anything.
This is the discipline most business owners skip. If you don’t write down your hypothesis, you’ll rewrite history after the fact and convince yourself you were right all along.
Run one change at a time, long enough to see whether it actually worked. Not 48 hours, not “it feels better”. Give it weeks, not days, and accept that most of your ideas will fail.
That’s not a bug in the process – it’s the process. Every failed test removes an assumption. The losses are how you find the wins.
Finally, build the habit before you build the stack. You don’t need expensive testing software to start. You need a cadence. One hypothesis a month, properly tested, properly measured, is worth more than a hundred gut-feel tweaks.
CRO fails because people treat it like a haircut when it’s actually a gym membership. Show up every week and the results will take care of themselves, much like how a consistent workout routine leads to noticeable improvements over time.
Consistency is key.
Alon Abraham notes that small businesses should focus on making small changes, more often, with a structured way of deciding what works and what doesn’t, rather than trying to make large changes all at once.
They should start with a question, not a tactic, and write down their hypothesis before making any changes. This will help them to avoid rewriting history after the fact and convince themselves they were right all along.
Small businesses should also focus on workers’ pay and how it affects their business, as it can have a significant impact on their bottom line.
By following these steps and being consistent, small businesses can improve their CRO and see noticeable improvements over time.
