GEO Measurement: Invest Anyway.
The most common objection to GEO investment is attribution — "we can't measure it precisely yet." True, and insufficient. The firms waiting for perfect measurement are ceding citation share that compounds over the next 18 months. Here is the case for measuring what you can and moving.
In every diligence conversation with a firm new to GEO, the same question surfaces within twenty minutes: "how will we know it's working?" The honest answer — the one an operator in a related discipline will recognize — is that the measurement is imperfect, improving quickly, and already good enough to make decisions on.
This is the same situation brand marketing has occupied for fifty years. It did not stop firms from investing.
What we can measure
These are not proxies. They are direct measurements of visibility in the layer that matters. You can run them today, re-run them in 90 days, and observe whether the interventions moved the number. That is more measurement than brand-tracking studies have offered most marketers for decades.
What we can't measure yet
- Last-click attribution from an AI conversation to a prospect arriving at your site — many AI platforms strip or minimize referrer data
- Whether a given prospect arriving via word-of-mouth has been influenced by an AI conversation they had in between
- The exact prompt the prospect ran, in most cases
- Cross-session influence — the prospect asked ChatGPT three weeks ago, is only now reaching out
These are real gaps. They will close. Platforms are adding referrer signals. Analytics vendors are building AI-conversation attribution models. The gap between "measurable" and "the real causal chain" is shrinking quarterly.
In the meantime, we do what every other mature discipline does: we measure what we can, we triangulate with adjacent signals (direct traffic patterns, branded-search volume, self-reported prospect sources), and we make decisions.
The cost of waiting
The argument for waiting — "we'll invest once attribution is clean" — sounds prudent and is, in fact, a very expensive decision. Here's why.
Citation share compounds. The firms getting cited in Q1 2026 are disproportionately the ones AI systems will default to in 2027 and beyond, because model training data and live retrieval both reward existing associations. The firms late to citation share will not just have missed the first movers' advantage — they will be fighting against an entrenched default.
This dynamic is specific to AI-layer competition. In Google, a new page can climb to rank one in a matter of months with enough authority behind it. In AI, a firm's category association is learned across many surfaces and reinforced over time. Catching up is harder than getting there first.
A way to think about the investment case. Ask the same question a brand marketer would have asked in 1985: is the category in which you compete going to be mediated through this channel in five years? If yes — and for professional services in the NYC tri-state area, the answer is already yes — then waiting for perfect measurement is a decision to arrive late.
How we recommend phasing the investment
Measure first
Start with a baseline. Understand the gap in your category. Understand which competitors are benefiting from your absence. The diagnosis is cheap relative to the decision it informs.
Execute the highest-leverage third of the roadmap
Not the whole thing. The highest-leverage third — the interventions that touch third-party citation presence, credential verifiability, and one or two core pieces of definitional content. That's what the first six months should be.
Re-measure at 90 days, then quarterly
The point is not to hit an absolute target; it is to observe whether the direction is right and whether the pace is acceptable. Adjust priorities based on what moved and what didn't.
Attribution is not the only form of evidence
Clean last-click attribution is a rare luxury, not a universal standard. The firms that invested in brand during the periods when brand attribution was weakest are the ones that built the enduring category positions. The firms waiting for GEO attribution to be airtight are replaying the same conservative mistake in a new medium.
Measure what you can. Invest anyway. Re-measure. Adjust. The firms that do will be the defaults when the category matures. The ones that don't will spend the next five years trying to unseat them.