Week 5 · Scaling, Optimisation & Handoff

Backed by App\Support\NamedExperts::get(). --}}

What you'll learn this week

By now you have an account built on Week 1's auction mechanics, an intent-bucket map from Week 2, CPC bands and a bid strategy from Week 3, and a creative testing queue running from Week 4. What's still missing is the operating rhythm that keeps all four of those working together month after month without a strategist manually rebuilding the account from scratch every few weeks. That rhythm is this week's subject.

Specifically, you'll learn the monthly optimisation cadence we run on every healthcare account past its first six weeks, the concrete triggers that tell you an account has outgrown its current tier — Starter, Growth, or Scale — and what a clean handoff between tiers actually looks like on both the agency side and the client side, because a badly run handoff is one of the most common ways a well-built bucket map quietly degrades. You'll close the week with an exercise that has you audit your own account's current tier fit and draft the first version of your own monthly optimisation calendar.

The prerequisite for this week is everything before it. If your bucket map from Week 2 hasn't been recalibrated at least once, or if you haven't run at least one creative test to completion from Week 4, some of what follows about cadence will feel abstract rather than concrete — go back and close those loops first if you can, though reading ahead won't hurt you.

Core concept 1 — the monthly optimisation cadence

Every healthcare account we run past its first six weeks settles into the same underlying rhythm, even though the specifics differ by tier. There's a mid-month checkpoint and an end-of-month review, and the two do different jobs. Understanding why they're separate, rather than folding everything into one monthly meeting, is worth spending real time on, because collapsing them is one of the more common ways teams let an account drift.

The mid-month checkpoint is a diagnostic, not a decision point. Its only job is to look at bucket-level cost per qualified lead across your intent stages — the same four stages you built in Week 2 — and flag anything that's moved meaningfully outside its normal band. A conversion-ready bucket that's suddenly costing forty percent more than its trailing average is worth flagging on day fifteen, not waiting to discover on day thirty when a full month of overspend has already happened. Nothing gets changed at the mid-month checkpoint beyond urgent corrections — a bid that's clearly broken, a creative variant that's clearly underperforming and burning budget. Everything else waits for the end-of-month review, because reacting to two weeks of data on a bucket map built for monthly patterns tends to introduce more noise than it removes.

The end-of-month review is where the actual recalibration happens, and it follows a fixed sequence rather than an open-ended discussion. First, the bucket-level CPC bands from Week 3 get checked against the month's real auction data and adjusted where the gap between projected and actual has grown beyond a small tolerance — usually somewhere around ten to fifteen percent, though the right threshold depends on how much budget the bucket carries. Second, the creative testing queue from Week 4 gets reviewed: winning variants get promoted to the default, losing variants get retired, and the next month's test slate gets drafted. Third, and this is the step teams skip most often, the bucket map itself gets a vocabulary check — are the representative query phrases from Week 2 still the ones actually showing up in the data, or has the specialty's conversational language shifted enough that the template needs updating.

One pattern worth naming explicitly: the monthly cadence is not the same thing as monthly reporting. Reporting is a summary of what happened. The cadence is a working session that changes what happens next. An account can have excellent monthly reports and still be running on stale bucket maps and untouched CPC bands, because nobody ever converted the report into a recalibration decision. Build the cadence as a decision-making session with a fixed agenda, not a read-out, and the reporting will naturally improve alongside it because there's more worth reporting.

Across the accounts we manage, this cadence is also where specialty and city priorities shift. A specialty that was the account's strongest performer in month two can slip by month five as auction competition in that specialty grows, while a specialty that started weak can strengthen once its bucket map has had three or four recalibration cycles to mature. The monthly review is the mechanism that catches that shift and reallocates budget accordingly, rather than leaving spend anchored to whichever specialty looked best when the account first launched.

Core concept 2 — tier upgrade triggers: Starter to Growth to Scale

Tier upgrades get treated too often as a budget conversation alone, and that's a mistake that leads to accounts moving up a tier before they're structurally ready for it, or staying stuck at a tier long after they've outgrown it. The triggers we actually watch combine budget signal with bucket-map maturity, and the two need to align before an upgrade makes sense.

The move from Starter to Growth has two triggers. The first is budget saturation: the account is consistently spending its full monthly allocation before the month ends, and — this qualifier matters — conversion-ready queries specifically are still losing auctions for lack of bid room even after CPC bands have been recalibrated at least once. Saturation on its own, without that second condition, sometimes just means the account needs a CPC band adjustment within its current budget rather than a tier change. The second trigger is bucket-map maturity: the intent-stage map from Week 2 has stabilised, meaning month-over-month recalibrations are producing small adjustments rather than large ones, which signals the account has learned what it needs to learn at its current scope and is ready to support more specialties, more cities, or a standing creative-testing programme wider than Starter is built to run.

Growth accounts differ from Starter in scope and in team structure, not in method — everything from Weeks 2 through 4 still applies, just run across more bucket maps at once. Where a Starter account typically runs one specialty-city combination, a Growth account runs several concurrently, each with its own bucket map, CPC bands, and creative queue. The team allocation changes too: Growth accounts get a named strategist plus a dedicated analyst who reviews bucket performance weekly rather than only at the monthly cadence, because more concurrent bucket maps means more surface area for something to drift unnoticed between reviews.

The move from Growth to Scale is more organisational than budget-driven, and this is where teams most often misjudge readiness. Budget saturation at engagement is a necessary condition — the account needs to be genuinely constrained by its current scope — but it isn't sufficient on its own. We look for three things together: the account is running five or more specialty-city bucket maps concurrently, the client's internal team needs direct dashboard access into bucket-level performance rather than relying on monthly reporting, and conversation-completion attribution data needs to start feeding into a broader marketing-mix model alongside the client's other channels rather than standing alone. An account can be heavily saturated at engagement and still not be ready for Scale if that third condition — the organisational readiness to consume Scale-level reporting and integrate it — isn't there yet on the client side.

A tier upgrade never means starting the bucket map over — the maps, bands, and vocabulary templates built at the lower tier carry forward directly. What changes at each step is scope, team depth, and reporting resolution, not the underlying method you've been building since Week 2.

Core concept 3 — running a clean handoff between tiers

A tier upgrade is also, almost always, a team change — a new strategist, sometimes a new analyst, occasionally a new reporting lead. The quality of that handoff determines whether the account's accumulated bucket-map history survives the transition or gets quietly lost and relearned the hard way over the following months.

On the agency side, the outgoing strategist runs a formal handoff session with the incoming team that covers every active bucket map, its full recalibration history, the current state of the creative testing queue, and any open questions or half-finished experiments. This isn't a document dump — it's a working session where the incoming strategist can ask why a particular CPC band sits where it does, or why one specialty's bucket map diverged from the generic template in a specific way, and get an answer grounded in that account's actual history rather than reconstructing the reasoning from raw data alone.

On the client side, we ask for one point of contact who can confirm business-side changes before they land in the newly scoped account — new specialties being added, new cities opening, pricing or service changes that would affect how conversion-ready queries should be interpreted. A handoff run without that confirmation step tends to import stale assumptions into the new tier: the incoming team builds on what the account currently shows without knowing that, say, a new specialty is about to be added next month and the bucket-map structure should be planned with room for it from day one.

Budget one to two weeks for a clean handoff, and run it alongside the outgoing month's normal optimisation cadence rather than instead of it — pausing the monthly cadence to focus purely on handoff logistics is how accounts end up with a month-long gap in recalibration right at the point when scope is expanding and the account most needs steady attention. Rushing a handoff to close within days almost always means skipping the documentation step, and an account that loses its bucket-map history at a tier change ends up relearning things Weeks 2 through 4 already established, which is a genuinely wasteful way to spend the extra budget a tier upgrade is meant to unlock.

Worked example — a real ICG engagement pattern (anonymised)

One of our diagnostic-chain engagements is a clean illustration of how the tier progression is meant to work when the triggers are read correctly rather than rushed. The account started at engagement with a single specialty-city bucket map — general diagnostics in one metro. Within the first two months, following the Week 1 through 4 process, the bucket map matured to the point where month-over-month recalibrations were producing small, stable adjustments rather than large swings, and the account was consistently spending its full Starter allocation with conversion-ready queries still losing a noticeable share of auctions.

Both triggers aligned — budget saturation plus bucket-map maturity — and the account moved to Growth in month three, expanding to three specialty-city combinations as the client added two more diagnostic categories in the same metro. The handoff followed the two-week pattern described above, with the outgoing engagement strategist walking the incoming engagement team through the original bucket map's full recalibration history before the new specialties were added on top of it, rather than building the new bucket maps in isolation from the established one.

The account has stayed at engagement for several months since, deliberately. Budget saturation reappeared around month five, but the second condition for a Scale move — organisational readiness on the client side for direct dashboard access and attribution data feeding a broader marketing-mix model — wasn't there yet, because the client's internal team was still building out the reporting infrastructure that would consume engagement data usefully. Rather than upgrading on budget signal alone, we recalibrated CPC bands within engagement and accepted that some conversion-ready volume would go unserved during peak periods, which the client's own unit economics supported. That's a legitimate outcome of a correctly read trigger set — not every saturated account should scale, and this one is a case where staying put was the right call, at least for now.

Exercise for you to do this week

This exercise doesn't need to end in a decision. Its value is in making the two triggers — budget saturation and bucket-map maturity — concrete against your own data, so that whenever the real decision does come up, you're reading it from evidence rather than instinct.

Common failure modes to avoid

• Treating budget saturation as the only upgrade trigger. Saturation without bucket-map maturity, or without the client-side organisational readiness a Scale move requires, usually means a band adjustment is the right fix, not a tier change.
• Collapsing the mid-month checkpoint and end-of-month review into one meeting. The two do different jobs — early flagging versus full recalibration — and merging them tends to produce either premature reactions to two weeks of noisy data or a full month of drift going unflagged until it's already expensive.
• Rushing a tier handoff to close within days. A handoff done in a hurry skips the documentation step almost every time, and the incoming team ends up relearning bucket-map history that already existed, wasting exactly the budget the upgrade was meant to unlock.
• Letting monthly reporting substitute for the monthly cadence. A report summarises what happened; the cadence is where recalibration decisions actually get made. An account can look well-reported and still be running on stale, unrecalibrated bucket maps underneath.
• Building new bucket maps in isolation after a tier upgrade instead of extending the existing history. New specialties or cities added at Growth or Scale should build on the recalibration history already established, not start from a blank Week 2 exercise as if the account were new.

Related specialty Dental Clinics Chair-count-aware conversational bidding for dental chains and solo practices. Dermatology Cosmetic + medical derm split campaigns with ASCI-safe claim language. Cardiology High-intent cardiac consult capture with NMC-clean claim language.

Compliance NMC Section 6 Doctor and clinic promotion rules applied to AI-assistant placements. DPDP Act 2023 Consent-flow and data-handling rules for conversational lead capture.

Case scenarios Multi-Specialty Hospital, North India Scale-tier rollout across a multi-specialty hospital network. Dermatology Clinic, Single Location Starter-tier conversational bidding for a solo dermatology practice.

Glossary ChatGPT Ads What conversational advertising inside ChatGPT actually means. Intent-Stage Bidding Bidding by conversation stage instead of by keyword. AI Assistant Channel (GA4) Tracking ChatGPT-sourced traffic as its own GA4 channel.

Exploring ChatGPT Ads for healthcare? Visit the ChatGPT Ads pillar →