From One Performance Max to Six: Scaling Sugar Babies After the Rebuild

The Setup

In early 2025, I rebuilt the Sugar Babies Google Ads account around three structural fixes: pulling brand out of a do-everything Performance Max campaign, rebuilding the product feed with custom labels, and repairing server-side tracking that had broken during the Shopify checkout-extensibility migration. The rebuild lifted blended ROAS 1.9x and grew non-brand revenue 147% over 90 days. Full story here.

Fourteen months later, the brand was ready for the next layer. The rebuild had bought a clean foundation. The next move was architecture: how does a mixed-catalog account with a hyperlocal Pacific Northwest pocket and a high-AOV furniture tier scale through Google Ads without overwhelming Smart Bidding’s signal?

What I Did Next

Expanded one Performance Max campaign into a six-campaign matrix. Two geography splits crossed with three value-band tiers. US National crossed with Mid ($75-$500), Premium ($500+), and Bundles. Hyperlocal Washington crossed with the same three tiers. Six campaigns running in parallel, each with its own budget, creative, audience seeds, and search themes. The matrix carries the bulk of paid acquisition spend now and replaces a single campaign that had been averaging across signal classes too different to optimize together. Full breakdown of the matrix here.

Seeded 26 Customer Match segments into the bidding signal. Past purchasers, repeat purchasers, AOV-tier cohorts, geographic cohorts, recently engaged but not yet converted. Each segment built from Shopify customer data, uploaded to Google Ads via the Customer Match API, refreshed on a rolling cadence. The segments don’t get used as direct targeting (Performance Max decides that). They get used as quality signal. Smart Bidding sees “this kind of buyer is who converted in the past” and steers toward similar profiles. Each campaign in the matrix gets the segments that match its tier.

Refreshed all 300 search themes across the matrix. Each Performance Max campaign supports up to 50 search themes. Six campaigns means 300 total. The pre-matrix campaign had one generic set. The new matrix has six tier-specific sets: mid-tier national gets themes around fast-moving accessory queries; premium national gets themes around considered-purchase categories; bundles get themes around “nursery set”, “complete bedroom”, “room package” intent; hyperlocal campaigns get themes that include the state name plus product category. Each set was rebuilt from the queries that drove revenue over the prior 90 days.

Built the catalog SEO foundation the matrix runs on. Performance Max pulls signal from landing pages. If a buyer searches “organic cotton baby clothes” and lands on a collection page with empty meta and no body copy, the bid model has nothing to learn from. The catalog work shipped in parallel with the matrix build:

Layer	Volume shipped
Product image alt texts	12,334 added across the active catalog
Collection meta descriptions	55 hand-written for active collections
Collection body copy rewrites	24 top-priority collections in brand voice
New collections built	1, an “In Stock & Ready to Ship” smart collection (~2,300 products)
URL redirects	7 (3 page-level + 4 collection-level)

The collection rewrites matter more than they look. Twenty-four of the highest-traffic collection landing pages now have real body copy on them. That’s the surface Google reads when it scores landing-page relevance for the matrix.

Why This Sequence

Fourteen months out from the rebuild, the account had earned the right to scale. The foundation was clean. The feed was tagged. Tracking was within 4% of GA4. The structural move at this scale isn’t another optimization pass. It’s a different campaign architecture.

One Performance Max campaign that absorbed the whole catalog had been the right structure at the original spend level. At the new scale, the same structure was averaging across three buyer cohorts that needed their own Smart Bidding contexts: a high-velocity mid-tier accessory buyer, a longer-cycle premium furniture buyer, and a high-AOV bundle buyer who converts in weeks rather than days.

The matrix isolates those contexts. The Customer Match segments give each context a quality signal. The search themes steer each campaign toward the queries that match its tier. The catalog SEO foundation feeds Google’s landing-page engine the words it needs to read. None of these moves work alone. All four had to ship together.

The Takeaway

The hardest part of a rebuild is the foundation work: feeds, tracking, brand isolation, taxonomy. The hardest part of a scale-up is resisting the urge to keep optimizing the same architecture after it has outgrown its limits. The right move at month 14 was not another round of tuning. It was a new structure, built on the discipline the first rebuild had bought.

If your Google Ads account is past the point where one Performance Max campaign can carry it and you’re not sure what the next architecture should look like, that’s the conversation. The matrix is one option. Whether it’s the right one for your account depends on what’s underneath, which is what the STACK Audit verifies before any architecture move. The numbers the matrix produces for the next board update slot into the four-number board memo and the CAC payback curve.