Skip to content
Work
About
Blog
Talk to me

The business case

Is GEO worth investing in?

The forecasts being quoted are real. They are also forecasts. Here is what the published numbers say, where they conflict, and what would actually settle it for your business.

The short answer

The published case is strong on direction and weak on precision. Valuates put GEO services at roughly $886M in 2024 and forecast about $7.3B by 2031, a 34% compound rate. Measurements of how often AI answers appear vary from around 16% to over 60% of searches depending on who is counting. The decision should rest on your own sampled baseline rather than on either figure.

What ranked on 23 September 2026: The market-size figure traces to a Valuates report distributed through PR Newswire and repeated widely without its methodology. The AI Overview coverage numbers come from several trackers that disagree substantially, which is worth knowing before quoting one in a board paper.

What the numbers do and do not support

Both of the popular statistics are defensible when quoted properly and misleading when quoted flat.

  • Valuates: about $886M in 2024 to roughly $7.3B by 2031 at 34% a year. That is a services-market forecast, not a measure of returns anyone has earned
  • BrightEdge reported AI Overviews on around 48% of searches in 2026; Semrush measured about 16% in late 2025, Conductor about 25% in early 2026, and Xponent21 about 60% in April 2026
  • Coverage varies enormously by industry, with health and education consistently highest and retail lower
  • None of these tell you what share of your buyers use assistants, which is the number that decides your case
  • A forecast describes a market growing, not a guarantee that participants in it profit

Things that sound right and are not

  • ClaimAI Overviews appear on 48% of searches.

    One tracker's figure, quoted without the four others that disagree. Say who measured it and over what, or use a range.

  • ClaimThe market will be $7.3 billion, so the opportunity is proven.

    That is spending on services, which is a measure of how much is sold rather than how much it returns. Vendor spend growing is not the same as buyer results improving.

  • ClaimEveryone will have a GEO budget by 2028.

    Plausible and unevidenced. It is a reasonable bet, and it should be described as one.

Where to start

  1. 01Sample twenty buyer questions across the assistants and record whether you appear. That is your actual exposure, and it costs an hour.
  2. 02Ask your last twenty customers how they found and researched you, and count how many mention an assistant.
  3. 03Compare the cost of the cheap version of this work against what you already spend per acquisition on paid.
  4. 04Decide on your own two numbers rather than on a market forecast written for investors.
  5. 05Revisit quarterly, because this surface moves faster than any annual plan.

Questions

Are the GEO market forecasts reliable?
They are directionally consistent with what buyers report, and they are forecasts produced by firms that sell reports. Use them as evidence of direction, never as a projection of your own returns.
Which AI Overview statistic should I quote?
A range with the sources named, or your own sample of your own queries. Every single figure in circulation comes from a different methodology on a different query set.
What would change your mind about this being worth it?
Evidence that assistant answers do not influence purchases in a given category, or that a business's buyers simply do not use them. Both are testable by asking customers, which is why that question is on the list above.

Related

Doing this yourself and want a second opinion on what the assistants say about you? Tell me what you are seeing, or read the AEO Playbook for the whole framework.