24% more visits from AI answers in 71 days.

Collabstr is a marketplace where brands find, hire and pay influencers and content creators. It already did well in Google, on the back of thousands of programmatic pages and a set of free tools. AI engines were a different story: they named Collabstr when someone asked which platform to use, and rarely for anything a buyer asks after that. We ran SEO and GEO as one program with press and a YouTube channel. We started on July 15, 2026. In the first 71 days, visits from AI answers were already 24% above the full 90 days before we started, the revenue those visits brought in was 15% above, and revenue from organic Google was 6% above.

The smaller plan hit the bigger plan's target

Our proposal offered Collabstr three plans, and we wrote the growth targets for the middle one. Collabstr chose the smallest, with about 40% less output each month. We kept the middle plan's targets as the bar anyway. Visits from AI answers are the main measure of the engagement, and the target was 52,000 to 59,500 over 90 days. Collabstr reached 58,053 in 71, inside that range with three weeks of the term still to run.

The challenge

Collabstr handled its own SEO and had done it well. The site carried programmatic pages for creators by city and niche, a set of free tools, and comparison pages against the other platforms. Press ran separately, through contractors, one campaign at a time. What it did not have was one program tying the technical work, the content, the press and the brand together around the questions buyers now ask AI engines.

Named for picking a platform, missing after that.

We asked ChatGPT, Gemini, Claude and Perplexity 40 buyer questions, five times each, for 800 answers. Collabstr was named in 60% of answers about choosing a platform and 57% about finding creators. On questions about measuring the return on influencer marketing, it was named in none of the 100 answers. Across all 800 it ranked sixth by share of voice.

AI answers filed it under beginners.

Of the 221 answers that named Collabstr, 42% put small business, budget, beginner or self-serve wording in the same sentence. None described it as a fit for a large brand or an agency running many clients, which is where Collabstr wants to grow.

Most of the near misses in Google were not buyers.

We pulled every search where Collabstr sat just below the top results in Google. 93.5% of those impressions came from people trying to download a TikTok or Instagram video, who find the free tools and buy nothing. Working down that list by size would have spent the quarter on them, so we set it aside and worked on the buyer searches underneath.

Every site change goes through their engineers.

We have no publishing access to collabstr.com. Every technical fix arrives as a written spec that Collabstr's developers build in their own sprint, so each spec has to be precise enough to build without a follow-up call.

The approach

Each piece of work points at a gap the measurement found, and each one feeds the others. A data finding becomes a press story, the press story becomes an article, and the article links back into the pages Collabstr sells from.

Technical fixes, written as specs.

We audited the whole site first. The first spec fixed canonical tags, the line of code that tells Google which version of a page is the real one, and Collabstr's developers have shipped it. The next covers spam profiles created on the marketplace. We also wrote the author and trust pages, which tell Google and AI engines who writes each article and why they know the subject, and page speed work is under way.

New articles, first for search demand, then for AI answers.

Month one went after searches with real demand where Collabstr had no page yet, like Instagram influencer marketing, brand ambassador programs and influencer gifting. Month two moved to the questions AI answers left Collabstr out of, like earned media value and influencer marketing reporting. Where Collabstr's own booking data answers the question, the article is built on it, because that is the part no competitor can write. The articles link into the pages Collabstr sells from, like its payments and UGC-for-ads pages.

Press first, then the blog.

Collabstr's marketplace data goes to journalists before it goes on the blog, because a finding already published on the company's own site is old news to an editor. The first one, that YouTube gaming creators command 23% higher deal rates than Twitch creators, ran in PocketGamer.biz and Tubefilter. The blog version follows once the outlets have run theirs, and links back to them.

Two YouTube videos a month, with different jobs.

From month two, one video each month is built to grow the channel, like "How Much Should You Pay an Influencer?". The other takes a question where AI answers leave Collabstr out, like "How Do I Measure ROI on Influencer Marketing?". In our tests Gemini cited YouTube in 60% of its answers and the other three engines never did, so the measurement articles on the blog carry the same answers in text. Each long video is also cut into Shorts.

A free tool every month.

Collabstr already runs a set of free tools, like its influencer price calculator. Each month we write the spec for a new one aimed at buyers, starting with an earned media value calculator, which tells a brand what a creator campaign was worth in advertising dollars.

What shipped

The first two months of the term, as of September 23, 2026.