Improvement Programs — We Do Not Stop at the Number

Who actually fixes what the measurement found?

We do. A sprint turns the audit’s priority list into published assets and patched pages, each carrying a work order and a completion receipt. A program keeps that loop running month after month, with the reading and the moves on the same calendar.

We didn’t write a promise into the copy — we wrote verification into the contract: on every execution program, the before-and-after re-measurement is a contracted deliverable.

The four-week Refit Sprint, week by week

A concentrated push on the foundations: the pages, the technical gates and the evidence assets that decide whether an answer engine can use you as a source at all. The list below is what lands, not what gets worked on.

Week 1
Priority set and work orders The audit’s findings become a ranked move list, each move scoped to one asset with a target, a stated change and acceptance criteria.
Week 2
Technical gates and entity facts Crawler access, rendering and structured-data patches, plus an entity fact sheet that states names, aliases, domains and products consistently across surfaces.
Week 3
Core pages and evidence assets Priority pages implemented or patched and new evidence assets published — the material an answer engine can quote and attribute.
Week 4
Receipt ledger and matched remeasurement The change ledger closes with before/after captures, URLs and timestamps; the panel reruns under the sealed protocol and the delta memo goes to the decision table.

Search Console, Bing and GA4 evidence is connected during the sprint, so the closing read can be set beside numbers you can recompute in your own accounts.

What the eight-week Category Sprint adds

Two implementation waves instead of one, and an evaluation design that is registered before the data is collected.

Three measurement points

Baseline, mid and final

A mid-point read between the two waves, so the second wave is ordered by what the first one actually moved rather than by the original plan.

Registered evaluation

Held-out and delayed reads

Part of the panel is held out from the intervention and part is read on a delay, both registered before collection — so a claimed improvement is checked, not just asserted.

Authority sources

A source pack, plus weekly memos

A target source list chosen from measured citation data with approval-ready material, and a written decision memo every week on what to scale, hold or stop.

The Category Sprint can start without a prior paid audit: its own baseline read opens the engagement. Starting from an existing sealed baseline simply makes the first week shorter.

How do you confirm the work did something?

Three gates, in order. Each one exists because a claimed improvement is easy to assert and expensive to check — so the checking is written into the deliverable.

V1 · Baseline sealed first

The starting line goes on the record

Before work begins, your panel’s baseline is sealed into a hash ledger — queries, surfaces, repeated runs and judging rules fixed and receipted, so later movement is read against a record that cannot be quietly edited.

V2 · Changes carry receipts

Every executed move is receipted

Work orders, before-and-after captures, URLs and timestamps bind each change to the evidence behind it — the change ledger ships with the deliverable and stays yours.

V3 · Matched remeasurement

The same panel, rerun under the same protocol

The closing read reruns your panel under matched conditions and publishes the delta with its confidence intervals — and the statements behind our public copy sit on the public claims registry, so you can check the verdict without trusting us.

Why the diagnosis comes first

Without a sealed t0 there is nothing for the closing read to be compared against, and the engagement ends with an opinion instead of a delta. A paid audit on the same domain within the previous thirty days supplies that baseline — and its full list price is credited toward the program’s first payment.

The credit rules, the eligible programs and a calculator for your own combination are on the pricing page.

What the engagement covers

Implementation on assets you own or can grant access to: pages, profiles, structured data, crawler and rendering configuration, and published evidence material. Media buying, platform advertising and third-party publication fees are contracted separately.

Where measurement cannot reach — in-app search, voice, personalized screens — the work is carried in the execution layer and named as such on the coverage map.