Reading the sets
One of the quickest ways to exhaust yourself at Fistral is to mistake movement for progress.
You see it quite easily on a messy day. Somebody relatively new to surfing paddles at almost everything. A lump appears, they go for it. Whitewater appears, they go for that too. Twenty minutes later they are knackered, roughly where they started, and beginning to look personally offended by the Atlantic.
Further out, somebody with more experience may appear to be doing very little.
That is the bit people miss.
They are watching. Reading the sets. Letting the wrong waves pass and moving when there is actually something worth moving for.
Website optimisation should work much more like that.
A lot of it does not.
The urge to touch everything
When enquiries are slow, sales dip or somebody decides the website "isn't converting", there is an almost irresistible urge to start changing things.
Make the button brighter.
Move the form.
Add a pop-up.
Remove half the homepage because somebody on LinkedIn announced that nobody reads anymore.
Put the homepage back because somebody else announced that long-form copy is essential.
This can go on for quite some time. It is highly active. It can also be almost completely useless.
The problem is not that changing a button, form or piece of copy can never improve conversion. Of course it can. The problem is making changes before you understand what is actually happening.
If visitors are arriving with the wrong intent, a prettier button will not rescue you. If the right people are landing on a service page and leaving because the offer is vague, redesigning the footer is unlikely to be the breakthrough.
You have to know where the wave is breaking.
Start with intent, not interface
I am much more interested in what somebody is trying to do than whether a page looks superficially "optimised".
Where did they come from?
What did they search for?
Which page did they land on?
What did they look at next?
Where did they stop?
And, importantly, was that person ever realistically going to buy anything in the first place?
Traffic is not one uniform substance. Ten people actively looking for the thing you sell can be more useful than a thousand people who arrived because you accidentally ranked for something vaguely adjacent.
That is why I would rather understand a smaller amount of meaningful traffic properly than celebrate a large number on an analytics dashboard simply because it is large and therefore pleasing.
Small data is still data. It is just not magic.
This is where people can get a little carried away with testing.
Controlled testing has its place when there are enough observations to distinguish a meaningful difference from ordinary variation.
But on a small website, an apparent uplift over a short test does not automatically prove that the change caused it. Change a button colour for a week, see a better number and declare victory if you like.
It is still not science.
It is theatre with a spreadsheet.
That does not mean you should ignore the data you have.
It means using the evidence you actually have rather than pretending it is stronger than it is.
Analytics can show patterns. Search data can give you clues about intent. Session recordings can reveal obvious friction. Enquiry forms can show what people repeatedly misunderstand. Client calls can tell you why somebody nearly did not contact you.
A simple question such as "what almost stopped you?" can sometimes reveal more than another dashboard full of coloured charts.
None of these is perfect on its own. Together, they start to tell you where the useful swell is coming from.
Make the obvious things genuinely obvious
Once you have somebody with the right intent, the website should stop trying to be clever.
They should be able to work out, fairly quickly:
what you actually do;
whether it is relevant to them;
what happens next;
what sort of commitment is involved;
and how to contact, book or buy from you.
This sounds painfully basic because it is painfully basic.
And yet there are still service websites where the visitor has to decode three paragraphs of abstract positioning before discovering what the company sells. There are pricing pages that avoid any meaningful indication of price. There are contact forms that feel like applying for planning permission.
I am not suggesting every business needs a single enormous red button saying BUY THIS NOW.
I am suggesting the path should make sense.
The homepage is not automatically the most important page
The homepage gets far too much emotional attention.
It is the front door, so everybody wants it polished. Fair enough.
But somebody arriving from search, an advert, an email or a referral may never use the front door. They may land directly on a service page, product page or article because that is the page that answered the question they had.
If that is where serious prospects actually enter the site, that page deserves serious attention.
This is another version of reading the sets: position yourself where the useful activity is happening, not where convention says it ought to happen.
Then leave it alone long enough to learn something
This may be the hardest part.
If you change the headline on Monday, the form on Wednesday and the offer on Friday, you have learned almost nothing by the following Monday.
Iteration is good. I am very pro-iteration.
Random movement is not iteration.
Make a change because you have a reason. Record what you expected to happen. Give it enough time and enough observations to produce useful evidence. Then decide whether you were right.
Sometimes you will be wrong.
Excellent. That is still information.
Know when to paddle
Conversion rate optimisation is not really about squeezing people through a funnel. I dislike that way of thinking about it because human beings are not toothpaste.
It is about removing unnecessary friction between somebody who already has a need and the thing that might genuinely solve it.
That requires analytics.
It also requires judgement.
Read what people are doing. Understand why they came. Stop chasing every ripple of advice that crosses your screen.
Then, when the set actually lines up, paddle.