Discussion starter: an SEO measurement question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.
- SEO Analytics
How do you value non-converting research visits from organic in reporting without inflating SEO’s contribution?
Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.
1 reply
-
adminAdmin
Value research visits by what happens to those visitors later, not by assigning a notional value per visit. The credible methods are assisted-conversion and path analysis, cohort tracking of people who first arrived through research content, a small set of honestly valued micro-conversions, and, where possible, an incrementality test. The discipline that prevents inflation is reporting “influenced” separately from “attributed”, with definitions agreed with finance, and never adding the two together. A research visit that led nowhere is worth nothing; one that started a relationship that later converted is worth something, and the point of measurement is telling them apart.
What not to do
- Traffic value from cost-per-click. Multiplying organic visits by what the same clicks would cost in paid search produces an impressive number that is not revenue, and finance teams know it.
- Crediting every later conversion to the first organic visit. Other channels, and the customer’s own intent, did most of the work.
- Counting the same conversion in several channels’ reports and presenting the totals side by side.
Methods that hold up
Method What it shows Limitation Conversion path analysis How often organic research sessions appear before conversions Depends on identifying users across sessions; consent and cookie limits reduce coverage Data-driven attribution in GA4 A modelled share of credit for organic touchpoints A model, not a measurement; GA4 retired its rule-based models in 2023, leaving data-driven and last click Cohort tracking Return and conversion rates over 30 to 90 days for users who first arrived via research content, compared with other cohorts Selection bias: researchers may differ from other visitors Micro-conversions Newsletter sign-ups, account creation, tool use, saved items Only valuable if the value assigned is realistic and validated Incrementality test What happens to conversions when research content changes or is withheld Rarely feasible; hard to design for SEO Brand search lift Whether research content grows brand demand over time Many other influences on brand search A reporting structure that stays credible
- Attributed revenue: organic’s share under the organisation’s standard attribution model, reported like every other channel.
- Influenced revenue: revenue from users who had an organic research session in a defined window before converting, clearly labelled as overlapping with other channels.
- Cohort outcomes: conversion and retention of users acquired through research content, compared with users acquired through other entry points.
- Leading indicators: micro-conversions and returning visitors from research content.
Agree these definitions with finance before reporting them. When finance helped define “influenced”, it is much harder for anyone to dismiss it as inflation.
Valuing micro-conversions honestly
If you assign a value to a newsletter sign-up or account creation, base it on what those users actually go on to be worth: the conversion rate of subscribers in the following months multiplied by average order value or contract value, discounted for what they would have done anyway if you can estimate it. Revisit the value every six months. A micro-conversion value set once and never checked drifts away from reality.
Hypothetical example:
A home improvement retailer’s guides bring large numbers of visitors who rarely buy in the same session. Path analysis shows that a meaningful share of customers who bought kitchen products in the quarter had read a kitchen guide in the previous 60 days. Cohort tracking shows guide-first visitors return and convert at a lower rate than product-page visitors, but at a much higher rate than visitors from display advertising. The retailer reports guide-attributed revenue under last click, guide-influenced revenue as a separate, overlapping figure, and the cohort comparison, with a note from finance agreeing the definitions.
Where research content earns its keep without revenue
Some research content is worth keeping for reasons that do not show up as revenue: it earns links that support commercial pages, it answers questions support would otherwise handle, it builds brand familiarity, and it is cited by assistants. Report those benefits separately, with their own measures, such as referring domains, support deflection or AI citations, rather than trying to convert everything into money. Not every valuable thing needs a revenue figure, and forcing one is where inflation creeps in.
Watch for this:
Influenced revenue figures can be large, because many customers touch organic search at some point. Present them alongside the attributed figure and the channel overlap, or they will be read as additional revenue, and the credibility of the whole SEO report will suffer when someone adds up all the channels’ influenced totals.
Consent and identity limit what you can see
All the path-based methods depend on recognising the same person across visits, and that has become harder. Consent choices mean some visitors are never tracked, cookie lifetimes are limited in some browsers, and people switch devices between research and purchase. The result is that influenced figures are almost always undercounts, and the undercount is uneven: it is worse for long research cycles and for mobile-to-desktop journeys. Logged-in users, account creation and email sign-ups give you the most reliable cross-visit identity, which is one more reason micro-conversions matter. When you report influenced revenue, say how much of the audience the analysis can actually follow, so readers know the figure describes a sample rather than everyone.
When a test is possible
For businesses with distinct regions, content sections or product lines, you can sometimes test incrementality: pause investment in research content for one comparable area, or launch it in one and not another, and compare downstream conversions over a quarter. It is rarely clean, but even a rough test is more persuasive than a model. Choose the test where you can live with the downside if the content turns out to matter.
Report attributed and influenced separately with finance-approved definitions, use cohorts to show what research visitors become, value micro-conversions from real downstream data, and report non-revenue benefits in their own terms.
Need help with this on your own site?
Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.
Add a reply Cancel reply
Replies are open to approved members. Log in or apply to join.