Conversion Optimization

Funnel analysis, recorded session review, form and checkout fixes, and a test plan ranked by expected value, every change measured against a baseline captured before anything moves. We will tell you early if your traffic is too thin for formal testing to reach a conclusion, because below a few hundred conversions a month you are better served by fixing clear faults directly than by waiting on significance.

Conversion optimization has a reputation problem, earned by a decade of articles about button colours. Real conversion work is unglamorous and mostly consists of finding the specific place where people who wanted to buy could not, and removing it.

That place is rarely where anyone guesses. It is a phone field rejecting a valid number, a delivery charge appearing at step four, a form that clears itself on a validation error, a mobile layout where the submit button sits under a sticky bar.

So the work is evidence first. Session recordings, funnel data, and the search terms people typed on your own site before leaving. Once the cause is known the fix is usually small, and the measurement is against a baseline captured before anything changed.

Losses found, not guessed at
Fixes measured against a baseline
Revenue per session, reported honestly

(Included)

06 items

Funnel analysis

Where sessions leave, segmented by device and traffic source, with each drop-off given a rough revenue value so priority is not a matter of opinion.

Session review

Watching real sessions at the failing steps. Ten recordings routinely explain what a month of aggregate charts could not.

Form and checkout fixes

Validation, field count, error handling, autofill and mobile keyboard behaviour, which is where a surprising share of losses live.

Test plan

Hypotheses ranked by expected value and effort, each with a stated sample size, so a test is not called the moment it looks favourable.

Speed work where it converts

Performance improvements targeted at the pages in the purchase path rather than at a homepage score nobody buys from.

Baseline reporting

What the numbers were before, what they are now, and what specifically changed between the two.

(How it runs)

01Establish the baseline
Nothing changes in the first pass. Without a clean before, every later claim of improvement is a story rather than a measurement.
02Find the actual cause
Data plus recorded sessions plus, where possible, watching someone unfamiliar try to buy. The third of those is the most uncomfortable and the most useful.
03Fix, do not redesign
The smallest change that addresses the cause. Redesigning the page changes fifty variables and tells you nothing about which one mattered.
04Measure, then move on
Hold the test to its planned sample size, record the result honestly including the failures, and go to the next largest leak.

(Is this the right thing to buy)

Who it suits: Sites with enough traffic to see where sessions die, and a team willing to accept that some tests will fail. It pays back fastest when you are already buying traffic, because every improvement to the conversion rate improves the return on every rupee of ad spend at the same time.
When it is the wrong call: Below roughly a few hundred conversions a month, formal testing takes too long to reach significance and we will say so rather than run tests that cannot conclude. At that volume, fixing clear usability faults directly is the better use of the same budget, and we are happy to do only that.
What it costs: Either a one-off audit with a prioritised list of fixes, which is where most clients should start, or a monthly programme covering analysis, implementation and honest reporting including the tests that failed. The audit is a fixed price. The programme is priced by how many changes are implemented per month and whether we are building them or handing specifications to your developers.

(Questions we get asked)

Do we have enough traffic for A/B testing?

Many sites do not, and it is worth knowing that early. Below roughly a few hundred conversions a month, tests take too long to reach significance and you are better served by fixing clear usability faults directly. We will tell you which situation you are in before proposing a testing programme.

Is this just about the checkout?

No, though checkout is where losses are easiest to price. Conversion work covers the whole path: whether the product page answers the question that stops people buying, whether search returns anything useful, whether the mobile layout works one handed.

What if a test makes things worse?

Some will, and that is a normal cost of finding out. It gets reverted and reported as a negative result. An agency whose tests all succeed is either not testing or not telling you about the ones that did not.

What is included in Conversion Optimization