Results
Nothing published here yet.
The first engagement is running. Its 90-day re-measurement does not exist yet, so there is nothing honest to put on this page. When there is, it will appear here in the format below.
Why this page is empty rather than padded
A results page is the easiest page on any agency site to fill dishonestly. Logos of companies that bought a workshop once. Testimonial quotes with no number in them. Projected figures presented in the past tense.
We sell measurement. Padding this page would tell you exactly how much our measurement is worth, so it stays empty until it is not.
The publishing policy
Clients are never named. Non-disclosure is policy rather than a case-by-case negotiation, because the companies worth working with are the ones who will tell us the truth about their data, and they only do that when naming them is off the table.
Redacted results are weaker proof than named ones. We accept that trade and compensate for it by stating the method on every entry, so the figures can be argued with even when the company cannot be identified.
What every entry will carry
Same fields, same order, every time. Including the last one.
| Field | What it states |
|---|---|
| Category | The client's category, not the client |
| Baseline | Citation share at the start, per engine |
| Delta | Citation share after 90 days, measured identically |
| Query set | Size of the set and how it was constructed |
| Engines | Which engines, at how many runs per query |
| Method | Method version that produced both measurements |
| What we did | The work, specifically enough to be argued with |
| What we cannot attribute | Movement we will not claim credit for |
The template below is the shape of a published entry, with the figures zeroed because none have been measured. It is shown here so the format is checkable before there is anything in it.
Zeroed template. Not a result. Method v1.0.
Same content as the machine surface. Switch between them to see what a crawler receives.
[ ZEROED TEMPLATE. NOT A RESULT. ]
## Redacted result 01
Category: CRE underwriting and valuation software
Engagement: Build, 8 weeks
Baseline collected: 0000-00-00 (method v1.0)
Delta collected: 0000-00-00 (method v1.0)
Query set: 38 category queries, 12 problem queries
Engines: ChatGPT, Claude, Gemini, Perplexity, AI Overviews
Runs per query: 3
Citation share, baseline -> 90 days
ChatGPT 0.0% -> 0.0%
Claude 0.0% -> 0.0%
Gemini 0.0% -> 0.0%
Perplexity 0.0% -> 0.0%
Google AI Overviews 0.0% -> 0.0%
What we cannot attribute:
[ Movement coinciding with an engine update is reported and not claimed. ]The one result we can publish without permission
martenfield.com is in the same tracking pipeline we sell, against its own query set, measured from the day the site went live. We expect the early numbers to show we do not appear for most of it.
Those get published quarterly in research, starting from the baseline where we look bad. It is the most credible case study available to a firm with no case studies, and it is the only client whose permission we do not need.