The leak audit
Every entry point you buy traffic to, measured at the four positions above. You get the numbers and the order to fix them in, which is not usually the order you expected.
Conversion rate optimisation
Most of the loss happens before the first question and after the last one
You have already paid for the click by the time any of this matters. Between the ad and the first answer, and again between the last answer and your sales team, there are four places where paid traffic quietly leaves. Your conversion rate averages all four and names none of them.
Not open yet. When it opens it rates your funnel on the four things that decide whether traffic converts.
The promise in the ad
They clicked a line about their problem and landed on a page about you. Most of the loss on paid traffic is here, and it looks like a bad keyword.
The first screen
Nothing on it says what they get or what it costs them in minutes. They are deciding whether to start, and almost nobody starts twice.
The run itself
Somewhere in the middle there is one screen that asks for more than it has earned. It sheds a fixed share of every campaign you run through it.
The handover
They finished, then met a form, or a thank-you page, or eleven days of silence. A lead that reaches nobody is spend you have already paid twice.
The problem
One conversion rate, four different failures underneath it
Two funnels can report the same rate and be broken in completely different places. The rate is identical either way, which is why it never tells you what to change.
In the first, most of the traffic arrives and never starts. That is a promise problem and it belongs to the ad. In the second, they start and stop halfway, so one screen is doing all the damage. In the third they finish, and hardly any of them reach a person, so the loss sits nowhere near the page you were about to redesign.
All three read as a low conversion rate. Only one of them gets fixed by better copy on the entry page, which is the first thing everyone changes.
So the work starts with four numbers instead of one. Started, finished, captured, routed. Where those sit tells you which of your money is going missing, and that decides what gets touched.
Before
One rate, measured at the end, explaining nothing.
After
Four rates, measured in order, and one of them is obviously the problem.
What is included
Six pieces of work, and every one of them is measured before and after
Not a list of tactics. Each item below is a position where paid traffic is lost, and the deliverable is the number at that position moving.
Message match, ad by ad
Your ad set against the page it lands on, line by line. Where the promise changes between the two, the spend behind it is already lost.
The entry page, rebuilt
What it says, what it asks first, and how fast it gets there. Written for the traffic that actually arrives rather than the visitor you hoped for.
The opener
The first thing you ask decides how many people answer anything at all. It is about them, it is easy, and it is never your budget or your job title.
The capture point
Where the email is asked for, what is said around it, and what is promised in return. Move it a screen earlier and you trade completion for volume, so we measure both.
What happens next
The result page, the report, and the first message out. Most funnels are optimised to the finish line and then stop, which is where the qualified ones go cold.
The design work behind the rebuilt pages is the same practice as funnel design. The difference is that this one starts from your numbers rather than from a blank page.
How it runs
Six weeks, and the first two are spent not changing anything
Rebuilding before you have a baseline means you never find out whether the rebuild worked. So the measurement comes first, even though it is the part nobody wants to pay for.
Measure what is there
Instrumentation on all four positions first, because a redesign with no baseline is a guess with a bill attached. If nothing is tracked yet, this is where that gets fixed.
Week 1
Find the expensive one
Rank the four leaks by the traffic behind each. The biggest percentage drop is rarely the biggest loss, because the cheapest channel is rarely the one carrying the budget.
Week 2
Change one thing at a time
Rebuilt entry page, reworded opener, moved capture point. Sequenced so you can tell which change did the work, rather than shipped together and argued about later.
Weeks 3 to 5
Read it back
The same four numbers, after. You get what moved, what did not, and the honest read on which changes were worth the disruption.
Week 6
Six weeks is where we would scope one entry point with live traffic behind it. More entry points and it runs longer. It is a judgement call, and the honest version is that the timeline depends on how fast your traffic produces enough numbers to read.
What you have at the end
Four things, and you keep all of them if you never speak to us again
- A ranked list of every leak found, with the traffic and the spend sitting behind each one.
- The rebuilt entry pages, live, in your account and under your control.
- A measurement plan naming which four numbers get watched from now on and who looks at them.
- A before-and-after read on each change, including the ones that did nothing.
The pages live in your account. The measurement plan is a page of definitions, not a licence. Nothing here stops working if the engagement does.
Packages
This sits inside two of the three packages
Separated by scope, not by how many hours anyone bills.
Package 1
One scorecard
Built and live, with the entry pages designed right the first time. A separate optimisation pass is not usually worth buying until traffic has run through it.
Package 2
The funnel around it
This work, in full. The entry pages, the capture point, the follow-up and the routing, treated as one thing rather than six.
Package 3
The programme
Everything above, plus someone watching those four numbers every month and changing whatever is not earning its place.
Prices go on the pricing page as numbers once they are fixed. Until then the scope is the honest part, and the scope is above.
What is in each packageWhere this is the first thing to fix
Four sectors where the spend is fine and the landing is not
These are the categories that already buy enough traffic for the four numbers to mean something within a month.
Vertical SaaS
You buy clicks from an industry that does not know your category exists. The entry page has to teach and qualify in the same breath.
Wholesale and distribution
Trade traffic arrives ready to price something and meets a brochure. The gap between those two is the whole leak.
Telecoms
High volume, low margin, and a landing page doing the same job as 40 competitors. Message match is worth more here than anywhere.
Manufacturing
Technical buyers, long lists of specifications, and a contact form at the end of it. The traffic is qualified and the page loses it anyway.
Questions
The five things people ask before they buy this
Timing and target. This is a project on the pages around the instrument, run before or shortly after launch, and it ends. The retainer is ongoing work on the instrument itself once responses are coming in. Most teams buy this once and the retainer later, and nobody needs both in the same month.
Your turn
Find out which of the four is costing you most
The scorecard rates your funnel on the four things that decide whether the traffic you buy turns into a call, and the report names the weakest one first.
Not open yet. The score and the report are yours whether or not we ever speak.