What happens between signing and the first memo.

Engagement letter countersigned, invoice issued

Lance sends a one-page engagement letter naming the category, the deliverable shape, the pricing, and the standard terms. You countersign. We invoice the kickoff ($4,500), due net 15. We do not require payment before measurement begins, but the queue starts moving on signing.

Your side: legal review (typically < 1 day for this length of document) + countersign. Our side: invoice + queue position locked.

Query-set scoping call (30 minutes)

One call with you, your designated team contact, and Lance. We bring a starting list of 15-20 queries we think represent how a buyer in your category actually searches. You bring your knowledge of which queries are commercially most important. We end the call with a frozen, hash-locked query set.

Same call: we lock the competitor cohort. Default for most categories is the 3-5 obvious named competitors plus 1-2 long-tail competitors you watch. Easy to adjust later.

Your side: one person who knows the category · 30 minutes. Our side: Lance · 30 minutes + 1 hour prep.

First baseline measurement run

Baseline measurement begins. Each of the locked queries runs in repeated full passes on each of the seven AI surfaces we measure: Perplexity, ChatGPT search, Gemini grounded, Microsoft Copilot via Bing, Google AI Overviews (the five citation-grade engines) plus Claude and Gemma (the two model-knowledge engines).

You receive a one-line confirmation email after each run (unless you opt out) saying the run completed with the OK rate. No memo content yet. The early runs are establishing baseline.

Your side: nothing required. Our side: scheduled cron + monitoring.

First-week status (optional check-in)

The first runs are captured. If anything looks structurally surprising (a competitor we did not include is dominating the cohort, the query set is producing answers that suggest a wording problem, an engine is rate-limiting), Lance flags it now rather than waiting for the memo.

Most engagements skip this check-in. It exists for the edge cases.

Your side: read a one-paragraph status email if it comes · 2 minutes. Our side: Lance · review the early data.

Continued baseline measurement

Same rhythm. Three clean runs is the minimum for the pattern-readiness rule to hold (see /methodology/). Each additional run adds to the signal-vs-noise resolution and the recurring-host classifier output.

Behind the scenes during this window: classification of every cited URL into the nine source-type buckets, competitive cohort tracking against the named competitors, per-engine behavior tracking, anomaly detection if any engine shifts its citation behavior mid-window.

Your side: nothing required. Our side: automated runs + monitoring + analysis prep.

Analysis and draft memo

Lance writes the research memo against the three weeks of captured data. Sections:

• Executive summary (the one thing that matters).
• Methodology and honest scope (what was and was not measured).
• The headline finding (specific to your data, not generic).
• Source-type distribution (which kinds of sources AI cites for your category).
• Per-engine breakdown (where the seven engines differ).
• Competitive gap analysis (which competitors appear on which queries).
• Recurring host analysis (the specific authorities AI engines learned to trust).
• A prioritized punch list you or your agency execute, sequenced by our estimate of leverage relative to implementation cost, with effort estimates and, for each item, the earliest window we could observe any movement, if it occurs.
• What we did not measure (honest scope limitations).
• Appendices (raw data pointers, classifier rules, query set).

See /example-engagement/ for what this looks like in finished form.

Your side: nothing required. Our side: Lance · ~6-10 hours of focused writing.

Internal grader review

The draft memo runs through the fail-closed factual-and-voice grader before any version reaches you. The grader is a separate AI process that holds the canonical fact list every customer-facing artifact must honor. If anything in the memo asserts something we cannot substantiate, the grader rejects the draft and Lance revises.

Your side: nothing required. Our side: automated grader + Lance revision pass.

Memo delivered

You receive the research memo as a PDF (and as markdown, if you prefer that format too) attached to a short cover email naming the headline finding and the highest-priority punch list item. Delivered to your designated recipients only, typically two to four people on your team.

Your raw measurement data (every captured API response, every cited URL with timestamps) is available at your customer data store and is exportable any time.

Your side: read + circulate to your execution team · ~30 minutes for the headline. Our side: Lance · sends.

Optional walkthrough call

If your team wants to discuss the memo synchronously, we schedule a 30-minute walkthrough. Lance walks through the prepped punch list, takes questions, names what we would prioritize and why. If your team is comfortable reading the memo cold, we skip this step.

Your side: optional · 30 minutes. Our side: Lance · 30 minutes.

Continued measurement, monthly delta memo

If you opted into the ongoing engagement at scoping, the $1,500/month retainer takes over. Measurement continues each month on the locked query set. A monthly delta memo arrives on the same calendar day each month, reporting what moved, what did not, and any punch list refreshes based on the new data.

The first delta memo lands roughly 30 days after the kickoff memo. The format is lighter than the kickoff memo: focused on the deltas, with the full punch list status table appended.

You can cancel the ongoing engagement at the end of any month, no claw-back.

Your side: read the monthly memo · ~20 minutes. Our side: automated measurement + Lance · ~3-4 hours per month per category.