Launch the GEO program with a defined initial scope, committed capacity, and a date for deciding what comes next. Use the first ninety days to put selected improvements into operation, establish dependable evidence, and learn which work deserves continuing resources.
The program should help suitable prospective clients understand the firm accurately and reach the appropriate professionals. Managing it therefore requires attention to the information being published, the answers being observed, the inquiries being received, and the resources needed to sustain the work.
20.1 Turn the Approved Backlog Into a Phased 90-Day Launch Plan
Start with the priorities, monitoring specification, and improvement backlog already developed. Choose an initial scope narrow enough to maintain and meaningful enough to support a real buying decision. Define the practice, service, intended client, relevant geography, and buyer questions. “Improve the public explanation of working-capital diagnostics for midmarket manufacturers” gives the team a usable assignment.
Select a small group of related improvements. A clarified service page may need supporting practitioner profiles, an explanation of the engagement process, and an accurate inquiry route. Complete the dependencies that make those improvements useful. Do not fill the quarter with unrelated tasks simply because they are easy to count.
Days 1–15 Commit the scope and resources
Confirm the sponsor, program owner, participating experts, spending limit, and protected time. Check that the baseline records are complete and that the collection specification can be rerun. Confirm that the inquiry owner knows how to record discovery and qualification. Identify any access problem, unsupported claim, or unresolved permission that affects the initial work.
Record the initial conditions before planned changes where practicable. Correct a consequential factual error when it is identified; preserving a baseline does not justify leaving misleading information public. Log the correction and explain its effect on subsequent comparisons.
Days 16–30 Complete the first set of improvements
Move the selected work through the publication process established in Chapter 19. Verify the live result, retain the approved evidence, and assign maintenance. For technical changes, record what was tested and what the test establishes. A page being accessible does not demonstrate that an assistant has retrieved or recommended it.
Google’s current guidance explicitly says that satisfying its requirements and best practices does not guarantee crawling, indexing, or serving. Set the launch deadline around work the firm can complete; observe external inclusion separately.
For example, a consulting firm can launch an improved diagnostic-service explanation and two substantive buyer guides while holding a proposed case study until its evidence and permissions are complete. The plan should identify a useful first release that does not depend on the most difficult asset.
Days 31–60 Establish repeatable delivery
Complete the next approved work, inspect the monitoring records, and resolve the bottleneck revealed by the first release. Check whether expert time, publishing support, and inquiry handling are sufficient. Adjust the workload to actual capacity.
Run a bounded experiment where there is a clear hypothesis and a measurement method capable of informing a decision. Keep the change record and comparison conditions established earlier in the book. A small program may learn more from one interpretable experiment than from simultaneous changes to every page.
Days 61–90 Decide the next phase
Prepare a management review covering completed work, information quality, observed AI representation, commercial evidence, costs, and remaining uncertainties. Recommend a defined next period: continue the scope, revise it, reduce an activity, pause, or expand selectively.
The decision should follow the evidence available at that point. If the firm’s engagements take months to develop, day ninety may be too early to judge revenue. It is still possible to assess whether the process works, the initial business rationale remains credible, and another funded period would resolve a useful question.
20.2 Establish Weekly Delivery Routines and Monthly and Quarterly Management Reviews
Give each review a distinct purpose. Use the same action register and evidence records across the program so people spend their time making decisions rather than reconstructing status.
Use the weekly review to complete work
The program owner and people responsible for current tasks review what can be completed, what is blocked, and what needs a decision. Confirm the next action, owner, and date for each exception. Check the coming week’s expert and implementation capacity before admitting more work.
A short meeting is usually sufficient if the register is current. Routine progress can be recorded asynchronously. Bring a practice leader into the discussion for a specific unresolved question. Consequential errors follow the correction process promptly rather than waiting for the weekly meeting.
Use the monthly review to interpret evidence
Prepare a concise report that separates three kinds of evidence:
- Operating evidence: Verified releases, unresolved errors, blocked work, maintenance completion, and actual time consumed.
- Representation evidence: Visibility, accuracy, suitability, and citations within the defined monitoring cases and conditions.
- Commercial evidence: Relevant inquiries, qualified opportunities, progression, accepted engagements, and revenue where observable.
Show collection failures, unreviewed records, sample sizes, and changes to the monitoring method. Keep different platforms, markets, and question groups identifiable. A higher aggregate score can result from adding easier cases; it should not automatically be presented as improvement in the original scope.
Inspect the saved answers behind consequential findings. An answer may describe the firm more prominently while overstating its services. That finding needs a different response from a correct description that reaches few of the monitored cases.
Close the monthly review with decisions on corrections, experiments, the next work batch, and any measurement repair. Record what remains unknown. Avoid turning incomplete commercial evidence into either a success claim or an automatic judgment of failure.
Use the quarterly review to allocate resources
The sponsor and relevant practice leaders assess whether the program still serves the firm’s business priorities. Compare its resource needs and useful contribution with other plausible uses of the same capacity, such as improving referral relationships or a neglected service page.
Approve the next scope, budget, staffing, and decision date. Make explicit which activities will stop or receive less effort to make room for new ones. Renewal of a supplier contract should follow this decision, with enough time to observe the contract’s notice requirements.
20.3 Maintain Accountability Across Managing Partners, Practice Leaders, Offices, and Billable Professionals
Assign authority over program decisions as carefully as responsibility for individual publications. The content owner in Chapter 19 manages an asset. The program owner manages the portfolio of work and brings unresolved choices to the people authorized to decide them.
Role | Program responsibility |
|---|---|
Managing partner or executive sponsor | Sets the business priority, commits resources, resolves competing demands, and decides continuation or expansion. |
Program owner | Maintains scope, budget, calendar, evidence, and the action register; recommends decisions. |
Practice leader | Confirms the service proposition and expert capacity; resolves substantive practice questions. |
Office or market leader | Verifies local capabilities, identity, service availability, and relevant qualifications. |
Participating professional | Completes the specific expertise and review commitments accepted in the work plan. |
Business development or intake owner | Records inquiry and opportunity evidence and follows the firm’s normal acceptance process. |
Make participation a workload commitment
Replace “partners will help when needed” with named assignments and reserved time. The practice leader should agree which professional will contribute, what they will deliver, and which existing commitment must be adjusted. Use the firm’s workload planning and internal project codes where available.
For a small launch, an illustrative weekly allocation is four hours for marketing and program coordination, two for professional input, one for website work, and one for measurement. Eight hours across thirteen weeks is 104 internal hours. These are work allocations, not separate jobs or a promise about publication volume. Significant setup work requires additional provision.
Recognize useful expert contributions, evidence checks, and corrections. Avoid rewarding professionals solely for articles produced or AI mentions observed. Those incentives can encourage claims or volume that the firm cannot responsibly support.
Make the economic commitment visible
Show one-time setup spending, recurring external costs, and internal hours by function. Use a consistent internal cost convention. Do not add salary cost and the full billing rate as if they were two independent costs of the same hour. If work displaces a billable engagement, explain the opportunity-cost assumption separately.
The program owner should be able to narrow routine work within the agreed mandate. Changes to professional claims, service scope, or publication permissions still go to the appropriate reviewer. Escalate conflicts with a concrete choice, such as delaying a launch, funding additional qualified review, or reducing its scope.
20.4 Adapt the Operating Model for Solo Practices, Boutiques, Regional Firms, and International Networks
Keep the essential decisions consistent across firm sizes while changing how the work is organized. A solo professional needs a manageable personal routine. A network needs shared methods and clear responsibility for what each member actually offers.
Solo practices
Start with one service and a small set of consequential client questions. Use one working register and protected time for review and maintenance. The professional can conduct a written monthly review without convening a meeting with themselves.
Buy help for defined tasks, such as editing, website implementation, or monitoring. Retain responsibility for professional substance and client acceptance. Before purchasing a package, estimate the review time it creates. A large content commitment can overwhelm a solo practice even when the supplier does all the drafting.
Boutique firms
Assign one program owner and a small group of available experts. Share website support, monitoring definitions, and the publishing process. Select the initial practice according to business relevance and readiness, with a named backup for essential work.
Agree how another practice gains a place in the next work period. Partners should be able to see the selection criteria and capacity constraints. Otherwise, the program can become a succession of personal requests that disrupt the agreed scope.
Regional and multi-office firms
Centralize reusable methods and technical infrastructure while giving local leaders responsibility for office facts and service availability. A shared brand does not make every claim true in every office. Maintain a record of which team can deliver each advertised service and where inquiries should go.
For example, a regional accounting firm can reserve essential correction and maintenance capacity during its busiest delivery period and schedule discretionary expansion afterward. If expert availability falls, reduce the planned workload. Commissioning unchecked drafts does not solve the capacity problem.
International firms and networks
Use a shared framework for definitions, evidence, and reporting, with local approval of language, service scope, and relevant jurisdictional statements. Identify the entity that provides the service and the relationship between network members accurately.
A network website should not imply that every member offers every service or that one member’s experience belongs to all others. Assign maintenance to the team capable of verifying the claim. Agree how a local change reaches the central website and any other authorized version.
20.5 Expand Into Practices, Industries, Locations, and Languages When Evidence and Capacity Justify It
Treat expansion as a new allocation of resources. Completing the first ninety days does not itself justify adding more practices or markets. The proposal should explain the client’s needs, what the firm can credibly offer, and why the next addition deserves capacity.
Require a defined case for the addition
Before approving an expansion, answer these questions:
- Which buyer needs or decides will the addition serve, and what evidence supports its importance?
- Can the firm provide the service as described, with appropriate expertise and availability?
- What useful public information or substantiated experience can the firm contribute?
- Who will own professional review, implementation, inquiry handling, and continuing maintenance?
- What will the addition cost, what other work will it displace, and what evidence will inform its review?
Use several forms of evidence. Relevant inquiries, client and referrer interviews, recurring misunderstandings, credible market knowledge, and observed AI answers can inform the proposal. A high visibility score is insufficient if the inquiries concern work the practice cannot undertake.
Resolve serious weaknesses in the existing process before enlarging it. Repeated missed reviews, unresolved material errors, or unusable measurement records are reasons to repair the operating model. Fund the additional maintenance load as part of expansion.
Add useful scope in a manageable increment
A first addition might extend a proven service explanation to one industry with genuinely different buyer questions. Another might cover a second office that provides the service through identified professionals. Explain the relevant difference in the material itself.
Google’s spam policies address doorway pages that funnel users through unhelpful intermediate destinations and large-scale production primarily intended to manipulate rankings. Multiple locations or translations are not automatically problematic. Build useful destinations for the intended audience and avoid producing near-identical pages merely to occupy more queries.
For a language addition, commission review by someone competent in both the language and the professional subject. Confirm that the terminology, service description, qualifications, and inquiry route fit the intended market. A faithful translation may be useful without adding different substantive claims, but it still needs ownership and updates.
Include the website implementation in the assignment, using the technical principles covered earlier. Check that readers can reach the appropriate version and that important qualifications survive translation. Link related versions in the maintenance register so a later correction is considered across them.
Evaluate the new scope separately
Define the new monitoring cases and baseline before assessing the addition. Preserve reporting for the original scope. Expanding the set of practices, countries, or questions changes what a combined measure represents; show that change explicitly.
In a hypothetical law firm, relevant inquiries support investigating a second jurisdiction. Before publishing an expanded offer, the responsible professionals confirm whether and how the firm may provide the service there. They approve local explanations and the inquiry route. The firm’s normal conflicts and matter-acceptance decisions still apply.
Approve one bounded addition with a spending limit, named owners, and a review date. Expand further when its evidence and operating performance support the next decision.
20.6 Track Platform Changes, Reassess Investment, and Discontinue Low-Value Activity
Assign one owner to monitor changes relevant to the experiences the program actually measures. Check official documentation before changing a website setting or accepting a supplier’s explanation of a performance movement.
Maintain a short official watchlist
Source | What to review |
|---|---|
Google Search Central documentation updates | Changes to search guidance, eligibility, measurement, and technical requirements. |
OpenAI crawler documentation and ChatGPT release notes | Crawler purposes and controls; changes to the monitored product experience. |
Anthropic crawler documentation and Claude release notes | Crawler purposes and controls; changes to the monitored product experience. |
Add official sources for other products in the monitoring specification, and review the monitoring supplier’s own collection and scoring changes. Keep distinct experiences, such as Google Search’s AI features and the Gemini app, separately identified.
Recheck access decisions by purpose. OpenAI documents independent search and training controls, while Anthropic distinguishes training, user-directed retrieval, and search crawlers. The website owner should verify the firm’s intended policy against each provider’s current documentation and implementation.
For every relevant change, record the affected assumption, evidence, owner, and disposition: no action, investigate, test, implement, or revise measurement. An announcement can warrant examination without establishing the cause of a change in the firm’s results.
Preserve comparability when measurement changes
Log changes to the observed product, model where identifiable, search mode, account context, location, prompts, collection method, and scoring definitions. If a material condition changes, annotate the series. Where feasible, run old and new methods over the same cases during an overlap period. If the earlier experience is unavailable, retain the history and establish a new baseline.
New official data can also change the value of a paid tool. Google announced dedicated Search Generative AI performance reports in June 2026 and recorded worldwide rollout in August. Its announcement describes URL impressions and dimensions such as pages and countries. These measures complement a firm’s answer monitoring; they do not establish every mention of the firm or attribute revenue.
Assess the incremental contribution of each tool against the updated measurement needs. Keep a paid capability when its usable evidence or saved effort justifies its cost. Revise reports when a measure changes rather than silently treating a new series as a continuation of the old one.
Continue only with a defined reason
At the investment review, distinguish evidence of poor value from a question that remains unresolved. A small sample or long buying cycle may leave commercial contribution uncertain. A further period can be justified when the business rationale remains credible, the cost is bounded, and the next work can resolve something useful.
State that learning question precisely. For example: “Over the next quarter, determine whether inquiries about this service reach the right practice and meet our qualification criteria.” Name the evidence, owner, budget, and decision date. An open-ended promise that results will eventually appear is insufficient.
Reduce or stop an activity when it no longer serves a priority, produces no usable learning, duplicates a cheaper capability, or requires review capacity the firm cannot supply. Stop methods that depend on fabricated evidence or misleading representation regardless of their reported visibility gains.
Discontinue carefully. Export the records the firm is entitled to retain, close supplier access as agreed, and transfer maintenance responsibility. Keep useful, accurate service information and accessibility improvements. Ending a monitoring subscription or content campaign does not require undoing work that continues to help prospective clients.
