“Peter, all this data analysis in your insights sounds great, but once you spot something, how do you know how to fix it?” — - redacted
That is a question I received on a call last week from a potential client.
And I am not going to lie.
As someone who has been working with Shopify stores for over 12 years it’s kind of embarrassing to admit..
But it completely stumped me.
Luckily though, not because I don’t know the answer. It was more because I realised I didn’t know how to explain the answer.
I also didn’t want to be that guy and say…
“Experience.”
Or:
“We’ve seen this before.”
Or:
“You get a feel for it.”
And yes, that is partly true.
We have a massive bank of past experiments, audits, research, failed tests, winning tests, customer insights, category patterns, and weird little Shopify problems that only make sense once you have seen them 100 times.
But still.
“Experience” is not a process.
“Gut feelzies” is not a strategy.
And “we just know where to hit the hammer” is not a very helpful answer.
Those types of explanations always irk me a bit.
Because it makes expertise sound like magic.
And good CRO should not feel like magic.
It should feel like diagnosis.
Anyways, it annoyed me so much that after that call, I went back and tried to map out how we actually move from insight to solution.
When we see a problem in the data, what are we really diagnosing?
And the answer I’ve landed on, for now, is something I’m calling: “The M Trilogy”
The basic idea is this:
Your website is not just a collection of pages.
It is a sales environment.
And that sales environment has three jobs.
It needs to:
- Say the right thing.
- Present the right options.
- Make the next step easy.
That gives us three areas to diagnose:
- Messaging.
- Merchandising.
- Mechanics.
And this is the part I wish I had explained on the call.
Because when a page is underperforming, most people jump straight to:
- “Let’s redesign it.”
- “Let’s add more reviews.”
- “Let’s make the button bigger.”
- “Let’s add a bundle.”
- “Let’s improve the copy.”
But those are fixes.
They are not diagnoses.
And a better starting point is…
What type of problem is this?
Because an underperforming product page could be a symptom of a variety of things.
It might be a Messaging problem.
The customer does not understand the product, believe the claims, see the difference, or feel confident enough to act.
For example, the page says:
“Made with premium technical fabric.”
But the customer is thinking:
“Will this shrink, crease, or feel horrible after I wash it?”
The brand is talking about the feature.
The customer needs the buying question answered.
That is Messaging.
Or it might be a Merchandising problem.
The product is clear, but the buying option is not.
The customer does not know whether to buy one product, the starter kit, the bundle, the refill, the subscription, or the “best value” option.
The site has products.
But it is making the customer build the buying decision alone.
A skincare brand might sell cleanser, serum, moisturiser and SPF.
All good products.
But a new customer does not know where to start.
So the fix might not be “better copy”.
It might be a clearer starter routine.
Or a better bundle.
Or product tiles that say “best for dry skin”, “best first purchase”, or “30-day routine”.
That is Merchandising.
Then there is Mechanics.
This is where the customer wants to move, but the website gets in it’s own way.
The size guide resets the selected size.
The colour swatch does not update the image.
The cart drawer is so full of upsells that the checkout button disappears on mobile.
The subscription selector is confusing.
The filters reload the page and throw the customer back to the top.
The customer has intent.
The site creates effort.
That is Mechanics.
And most of the time, the best findings sit in the overlap.
A review quote beside Add to Cart is Messaging, because it builds belief.
But it is also Mechanics, because its placement supports the action.
A bundle name is Messaging, because it explains value.
But it is also Merchandising, because it packages the buying option.
A cart upsell is Merchandising, because it encourages a bigger order.
But it is also Mechanics, because it can either help the purchase route or get in the way of it.
So instead of saying:
“The PDP needs work.”
You can say:
“The PDP has a Messaging + Mechanics issue. Visitors are opening reviews and delivery information before Add to Cart, but those answers sit too far below the buy box on mobile.”
That is a diagnosis.
The designer knows what problem they are solving.
The copywriter knows what doubt needs answering.
The strategist knows which lever they are pulling.
The client knows this is not just someone’s opinion dressed up as CRO.
And that is the real point.
Most CRO language is too vague.
And sh*t I am guilty of saying this stuff:
- “We need more trust.”
- “The offer needs to be stronger.”
- “The page needs to feel more premium.”
- “The UX needs improving.”
Maybe.
- But where?
- Which page?
- Which element?
- Which customer question?
- Which action?
“We need more trust” could mean:
- The claim has no proof.
- The return policy is hidden.
- The delivery message appears too late.
- The bundle saving is unclear.
- The checkout adds a surprise cost.
Those are all different problems.
And they need different fixes.
That is why jumping straight to best practices is dangerous.
“Add reviews near the buy box” might help.
Unless the real issue is that the customer does not know which product to choose.
“Add a sticky Add to Cart” might help.
Unless the real issue is that nobody believes the claim yet.
“Add a bundle” might help.
Unless the customer is already overwhelmed by choice.
Best practices are not always wrong.
They are just usually context-free.
And context is the job.
So, to answer the original question properly:
How do we know what to fix once we spot something?
We do not start with the fix.
We start by asking where the problem stems from.
- Is it Messaging - Do people understand it, believe it, and care?
- Is it Merchandising - Are we showing the right product, offer, bundle, price, or buying option?
- Is it Mechanics - Can people find it, choose it, add it, and buy it without the site getting in the way?
Knowing where to swing the hammer is not magic.
It is diagnosis.
And the better the diagnosis, the less random the fix becomes.
Chat soon,
Peter
P.S. If you are looking at your own site this week, pick one page and ask yourself this”
Is this page failing because customers do not understand, because they are not being shown the right buying option, or because the next step is harder than it needs to be?
That one question will usually get you further than another “best practice” checklist.
