White label SEO for agencies: what a partner has to be able to report now
Most white label SEO is a rankings dashboard with your logo on it. Google now reports AI answer impressions with no clicks, no CTR and no queries attached.
The white label SEO product most agencies are offered is a rankings dashboard with their logo in the corner. Keyword positions, a traffic chart, a monthly PDF. It is easy to resell, which is exactly why so many partners sell it.
The problem is that it answers a question your client has stopped asking on its own.
The short version
White label SEO means a partner does the work and your brand is on all of it. That has not changed. What has changed is the deliverable: a client who once wanted to know their position for a list of keywords now also wants to know whether an assistant names them when a buyer asks it a question. Those are two different measurements with two different blind spots, and a partner who only owns the first one is selling you half a product while calling it whole.
What changed, and when
Two things happened in the last few months and both are checkable.
Google now reports AI answers separately, and tells you almost nothing about them. Since June 2026 Search Console carries a Generative AI performance report covering AI Overviews and AI Mode. Googleâs own documentation is explicit about what it holds: impressions, grouped by page, country, date and device. No clicks. No click-through rate. No queries. And it is not everywhere yet, in Googleâs words: âWeâre rolling out this report to a subset of website owners, allowing for thorough testing before rolling it further.â
So the instrument confirms your client was shown inside an answer. It cannot tell you what the answer said, whether the client was recommended or merely listed, or who was named beside them.
And the answers are not equally stable. SISTRIX studied 82,619 prompts across 1,548,213 snapshots, six countries and three platforms over 17 weeks from December 2025 to April 2026. Weekly source replacement came out at 5% on Google AI Overviews, 56% on Google AI Mode and 74% on ChatGPT Search.
That spread is the finding. A single number for âhow volatile AI search isâ is always wrong by a factor of fifteen depending on which surface you mean. And underneath the churn there is a stable core: in the same study, 53% of AI Overview prompts saw not a single source change across the whole 17 weeks.
Two populations, then. Get into the stable core and the position holds for months. Land in the rotating carousel and you are renting. Which one a client is in is invisible without measurement, and that is the honest argument for the work rather than a reason to panic anybody.
Why the old reporting product cannot cover this
If the partnerâs product is a rankings dashboard, three things are now outside it.
- Position does not measure citation. A page can sit at 40 and be the source an assistant quotes. A page can sit at 3 and be quoted by nothing. Grading the first as a failure and the second as a success gets the client backwards.
- Impressions and clicks have come apart. When an answer is delivered above the result, a page holds its position, gets shown more and is clicked less. Read as totals, that looks like growth. Read as a ratio over time, the gap is the clearest signal available that it is happening.
- Google is one of four surfaces. Search Console sees nothing from ChatGPT, Perplexity or Claude. Whatever your partner reports from Google alone, it is silent on the rest.
None of this means rankings stopped mattering. It means a report built only on rankings has developed a hole, and the hole is where your clientâs question now lives.
What to ask a partner before you resell them
Six questions. The answers are more revealing than any case study.
- âShow me a report you would put my logo on.â Not a template. A real one, anonymised. If the AI section is a single score with no method behind it, ask what the score is made of and watch what happens.
- âHow do you measure citation, and how many samples?â Cited is a rate across a window, not a state on a single day. A one-day reading of a source set that can turn over 74% in a week is a coin flip with a chart on it.
- âWhat do you do when the engines get my client wrong?â Assistants state things about businesses that are out of date or simply false. Correcting that is an owner problem before it is a marketing one, and most partners have no answer because they have never looked.
- âWhat happens in month one that is not a report?â Measurement with no work attached is a subscription to being told things.
- âTell me about something that did not work.â A partner with no failures on record has either not been doing this long or is not measuring.
- âWill you ever contact my client?â The answer is no, in writing, before credentials move.
What we can show, and what we will not claim
We run this on our own properties before selling it, which is the only reason we have anything to show at all.
The strongest evidence we hold is a controlled comparison on a real commercial account, VanBaltic. Pages that received the work were compared against a control group on the same domain that did not. Over the measurement window the linked pages rose 111% in impressions against the controlâs 9%, and clicks moved +33% against the controlâs -20%. The control falling while the treated group rises is the part that matters, because seasonality cannot explain a divergence in opposite directions.
Those are ratios on purpose. We do not publish absolute traffic figures for our own or our clientsâ properties, and we do not publish growth multiples, because on a small base both flatter and neither is evidence.
We also run independent properties that earn citations in their own right, such as snowverdict.com, because there is a real ceiling on what content published on a companyâs own domain can achieve when assistants systematically prefer third-party sources. That ceiling is the thing most white label SEO offers never mention.
And the limits, plainly. Nobody can guarantee an assistant will recommend a business. The work is durable; the position is volatile. What stops when a retainer stops is the watching, not the asset.
How the white label side actually runs
The mechanics are the same as any of our white label work, and they are boring on purpose:
- NDA first, signed before we see a credential. It is our default rather than a request.
- No client contact, ever, in writing. Our only channel is you.
- Your brand on every surface, or no brand at all. Reports, audits, exports.
- One senior contact, not a queue. No telephone game between your client and the person doing the work.
- You set the price. We quote you partner rates and what you charge your client is your business.
If you are weighing a partner more generally, we wrote the vetting checklist for white label partners and a straight account of how the handoff works. For the production side rather than the search side, white label WordPress maintenance covers the recurring half.
Where to start
Ask your partner what they can currently tell you about a client that Search Console cannot. If the answer is a number with no method behind it, you already know what you are buying.
If you want that answered about one of your own clients before you commit to anything, tell us which client and which questions their buyers actually ask. That is the measurement, and it is the part that decides whether any of the rest is worth doing.