MappedSkills Logo

The published method

How we work, what we measure, and what we cannot tell you.

This page exists to be checked. It sets out how the work is done, what is measured at each point, what a measurement can and cannot support, and where the judgement calls are.

The section on what cannot be attributed is not a disclaimer at the bottom — it is a third of the reason this page is here.

Why these disciplines are one chain

Two observations sit underneath everything below.

The first: when a business is absent from AI answers, the reason is usually that those answers are assembled from directories, listings and third-party pages the business is not on. We measured that on ourselves — across 28 unbranded runs on four systems in one day, we appeared zero times, while directory sources appeared in 15 of them. The fix is mostly not on the website.

The second: traffic that does not become enquiries is one of four failures — the demand, the page, the enquiry path, or the measurement — and the same symptom appears in all four. Measurement is what tells you which one you have.

Both point the same way. The work at the top of the chain and the work at the bottom of it are connected, and the joins between them are where a result quietly disappears. That is an argument for the joins belonging to somebody.

It is not an argument that every engagement contains all of them. What a given piece of work covers is decided per client.

What we measure

Six points on one chain.

  1. Capture

    • Discovery and source

  2. Convert

    • Landing

    • Interaction

    • The enquiry

  3. Measure

    • The qualified enquiry

      Accountable up to here

    • Opportunity and revenue

    Your numbers, from your system.

The same six points, grouped. No stage is drawn larger than another, and nothing here is a count, a duration or a rate.
The six points measured on one chain, with the accountability boundary drawnOne continuous chain with six points in order: discovery and source, landing, interaction, the enquiry, the qualified enquiry, and opportunity and revenue. The stretch before discovery is dashed because buyer demand that never reaches the business is not measured. One enquiry arrives from an open, dashed mark with no upstream segment: it has no usable trail and it still counts. A rule crosses the chain after the qualified enquiry: everything before it is what this firm is accountable for. The segment beyond the rule is drawn as a double line because it is the client's own data from the client's own system, and it is reported as theirs. No point carries a count, a size, a share or a rate.buyer demand we cannot seediscovery and sourcelandinginteractionno usable trail— it still countsthe enquirythe qualified enquiryACCOUNTABLE UP TO HEREopportunity and revenueyour numbers, from your systemnothing here is a count, a size or a rate
Six points on one chain. The dashed stretch before the first point is the honesty device used everywhere on this site — a dashed line means unmeasured — and the open mark arriving at the enquiry with no segment behind it is an enquiry with no usable trail, which still counts. The rule after the qualified enquiry is the accountability boundary. Past it the line doubles, because what follows is your own data from your own system and it is reported as yours. Nothing on the drawing states a quantity, because no quantity here is measured.
Discovery and source
How someone first found the business, where that is knowable.
Landing
What they arrived on, and from what.
Interaction
What they did before enquiring, at the level of pages and actions rather than surveillance.
The enquiry
Recorded at the moment it is made, carrying its own source, rather than reconstructed later.
The qualified enquiry
Whether it is real, reachable, from a business context, with a stated need, and not a duplicate. This is the accountable number.
Opportunity and revenue
Only where the client supplies their own close data, reported as the client’s numbers from the client’s system.

Every one of those is reported against a starting point captured before the work begins. Where no starting point was captured, there is no before, and we say so rather than manufacturing one.

What we do not measure, and cannot

Six things this method does not produce.

Revenue we do not control. We do not set the price, build the product, staff the sales team, follow up the enquiry or close the deal. Revenue is reported from client data; it is never promised.

A rank in any AI system. There is no such number. Order changed in every repeated prompt we measured, so a position would be a reading of noise.

Anything about systems we cannot test. Claude and Grok returned no testable result in our measurement. We hold no baseline for either and make no claim about either.

Statistical significance below a real traffic threshold. Below it, the work is research, diagnosis and design, and we describe it that way rather than calling an observation a test.

Enquiries with no usable acquisition context. Direct entry, missing referrer, consent declined, and manually received enquiries — a phone call or an email — carry no first-touch attribution at all.

A before-and-after where no baseline was captured first. This applies to our own site as much as to a client’s.

we do not promise
  • revenue we do not control
  • a rank in any AI system
  • anything about systems we cannot test
  • significance below a real traffic threshold
  • a source for an enquiry that carries none
  • a before-and-after with no baseline behind it

What we cannot tell you about where an enquiry came from

Attribution is useful, incomplete, and unevenly incomplete.

Attribution means connecting an enquiry back to how the person first found the business and what they were looking at when they enquired.

The same chain, with the six kinds of arrival that carry no usable sourceThe same six-point chain, recessive. The enquiry point stays solid and the accountability boundary is thickened. Six kinds of arrival reach the enquiry point carrying no usable source, each drawn as an open dashed mark and each named: a person who researched on a phone and enquired on a laptop; a visitor who declined tracking consent; a link shared in a message, which arrives with no referrer; someone recommended, who saw the business offline or typed the address in; a phone call or an email, which carries no first touch at all; and revenue where the client supplies no close data. The six are drawn identically and none is smaller than the others: no proportion is shown, because none is measured. All six stay inside the count. A measurement gap never disqualifies a legitimate enquiry.discovery and sourcelandinginteractionthe enquiryALL SIX STAY IN THE COUNTresearched on a phone, enquired on a laptoptracking consent declined — correctlyshared in a message: no referrer at allrecommended, seen offline, or typed ina phone call or an email: no first touchrevenue, where no close data is suppliedA GAP IS A GAP, NOT ADISQUALIFICATIONno proportion is shown: none is measured
The same chain you have just read, in a changed condition: it recedes, the accountability boundary thickens, and what arrives carrying nothing becomes the subject. The six categories are drawn identically and at the same size on purpose. A figure showing “unattributed” at a plausible-looking size would be a fabricated statistic in graphic form, and no measured proportion exists. None of the six is a residual, an error or an empty state — each is a finding, each is a real enquiry, and each stays in the count.

A person who researches on their phone and enquires from a laptop appears as two different people. We cannot join them.

A visitor who declines tracking consent is not tracked. That is correct, and it means their enquiry arrives with no source.

A link shared in WhatsApp, an email or a private message arrives with no referrer and is recorded as direct. In this market that is a large category, not an edge case.

Someone who was recommended by a customer, saw the business somewhere offline, or simply typed the domain in, arrives with nothing to attribute.

Phone calls and emails carry no first-touch attribution at all. They are entered by a person, with the source recorded as unknown, and they are never quietly assigned to a channel.

Revenue attribution depends entirely on data the client supplies from their own system. Where it is not supplied, that part of the chain is empty and shown as empty.

A measurement gap never disqualifies a legitimate enquiry. Whether an enquiry counts and whether we can tell where it came from are two separate questions. An enquiry we could not trace is still a real enquiry, it stays in the count, and the gap is reported as a gap — because moving it out would make the measurement look tidier and the business smaller.

Where a person decides, not a process

Five places, named rather than hidden behind process language.

  • Which demand is worth capturing. Much of what looks like demand in this market is students, job-seekers and competitors.
  • Which of four failures a symptom indicates, when the evidence is partial.
  • Whether an enquiry is a real fit. The website can check that an enquiry is real, reachable and in context. It cannot judge what an opportunity is worth to you, and no form can. That is a conversation.
  • What a partial measurement means, and what it does not license us to say.
  • When to tell a client a channel is wrong for their business, including when we would be paid to run it.

Where the other capabilities sit in this

Three subordinate roles, stated at their real size.

Paid media is a demand source chosen on merit for a particular business, on the same measurement layer as everything else. It is not part of every engagement and it is not a positioning claim.

AI searchoccupies five subordinate roles: a delivery capability, part of the method, a diagnostic that explains absence, a reason to publish properly, and — because we measured ourselves and published the result — a piece of evidence. It is a supporting capability and not what the firm is.

Automation, after the enquiry exists, is limited to routing it to the right person, acknowledging it, and reminders around a booked call. It gets this sentence and no more, because that is genuinely all of it. We do not implement CRMs, build lifecycle programmes or run sales operations.

What we need from you

Four requirements, not four requests.

The work does not function without them, which is why they are stated this way.

  • Access. Analytics, search console, ad accounts, the site itself, and the systems where enquiries land.
  • Decisions, at a reasonable pace. Work that waits three weeks for an approval loses more than it looks like it does.
  • Product knowledge. What you sell, who buys it, what they ask before they buy, and what a good customer looks like. Nobody outside your business has that.
  • Capacity to follow up. An enquiry that is not answered is not a result, and it is the one part of the chain we do not control.

What we do not know

Visible, rather than a footnote.

Absolute search volume is unknown to us. No keyword tool has been available in this programme, so we quote no volume, no difficulty and no cost-per-click figure anywhere on this site. Where a figure would normally go, we write unknown.

We have no analytics baseline of our own yet. Which means we cannot show you a before-and-after of our own site, and we do not.

Organic search is a 12-month-plus horizon. That is our position from three independent lines of evidence. It is a horizon, not a schedule.

Our AI measurement is one round. One day, one location, English, signed out, four systems, two others untestable. One measurement is not a trend, and we will not describe a change until there is a second.

If you want this run on your business

The method is the same whether or not you hire us.

It is written so you can check it. If you want it applied to your situation, either route reaches the same place.

Tell us what you're trying to fixBook a call