The Google Ads Account Structure I Use for a New Ecommerce Client

By Ralph West  ·  October 21, 2026

This is for a store launching its first real Google Ads account. You have a product feed, a handful of best sellers, and no conversion history yet. The single most important decision is how you organize campaigns around product economics, not product category. Everything else in the account structure follows from that one call.

Decide this first

Structure your google ads account structure around margin and price point, not around how your website navigation is organized. Your site nav exists for shoppers. Your account structure exists for Google's bidding algorithm and for your own sanity when you're reading reports at 11pm.

Group products by how much profit they generate and how expensive they are to acquire. A $180 jacket with 60% margin behaves nothing like a $24 accessory with 35% margin in an auction. If you lump them into one campaign, the algorithm will chase volume on the cheap item and starve the one that actually makes you money. I learned this the hard way scaling a DTC brand from $100K to $3M. We had one Shopping campaign with everything in it for the first four months. Our blended ROAS looked fine. Our profit did not.

What to look for

Campaign split by funnel stage and intent

For a new account I run three campaign types from day one:

• Performance Max for broad reach, built on your top 20-30% of SKUs by margin and proven demand (even pre-launch, use whatever sales data you have)
• Standard Shopping campaign for manual control over your hero products, where you want to see search term data
• Branded Search campaign, separate and small, to protect your name and capture people who already know you

Keep branded spend under 5-10% of total budget. It converts well but it inflates your blended ROAS and hides how your prospecting campaigns actually perform. Report on it separately, always.

Ad group structure by price tier, not by product type

Inside your Standard Shopping campaign, break ad groups into price bands:

• Ad group 1: products under $30
• Ad group 2: products $30-$75
• Ad group 3: products over $75

This lets you set different target ROAS or bid caps for each band. A $200 product needs a different acquisition cost ceiling than a $20 one. If you organize ad groups by "tops" and "bottoms" instead, you're mixing a $19 tee with a $140 coat in the same bid strategy, and the bidding system can't optimize for either one properly.

Single product group isolation for your top 10 SKUs

Take your 5-10 best sellers, or your best guess at them if this is pre-launch, and give each one its own product group inside its ad group. Everything else can sit in a catchall "all other products" group.

Why this matters: Google reports performance at the product group level. If your top seller is buried inside a group of 40 other items, you cannot see its real CPC, its real conversion rate, or whether it's worth a bid increase. Isolating it costs you nothing and gives you data you'll use in week two.

What to ignore

Skip Smart Shopping entirely, it's deprecated and folded into Performance Max now, so this is moot for new accounts, but I mention it because old blog posts still recommend it.

Ignore audience signals in Performance Max for the first two weeks. You don't have a pixel with enough data yet. Feeding in guessed lookalike audiences on day one just confuses the algorithm during its learning phase. Let it learn from the product feed and broad signals first, then layer in your customer list once you have 100+ purchases logged.

Ignore Display campaigns as a standalone line item in a new ecommerce account. The CTR and conversion rates are so low relative to Search and Shopping that it's rarely worth a separate campaign and budget line in month one. Performance Max already covers Display placements inside its own mix.

Don't build out dozens of ad groups by color or by size variant. Google already handles that at the item level inside your feed. Adding "Blue Jacket" and "Red Jacket" as separate ad groups just fragments your data for no benefit.

Common mistakes

The most common mistake is launching with one campaign and one ad group containing the entire catalog, then checking back in three weeks wondering why ROAS is mediocre across the board. The algorithm needs structure to learn efficiently. A flat structure forces it to average performance across wildly different products, and averages hide the winners and the losers equally well.

The second mistake is setting a single target ROAS for the whole account on day one. New accounts don't have enough conversion data for Target ROAS bidding to work well. I run Maximize Conversion Value with no target for the first two to three weeks, or until the campaign hits 30-50 conversions, whichever comes first. Then I layer in a target ROAS based on what the data actually showed, not what I hoped for in the media plan.

The third mistake, and I see this constantly with new ecommerce clients, is changing the account structure every few days because early numbers look rough. Shopping campaigns need time to exit the learning phase, usually 7-14 days of stable settings. If you restructure ad groups, change budgets by more than 20%, or swap bid strategies every three days, you reset that learning window every time. I've seen founders do this out of anxiety and then wonder why performance never stabilizes. Pick a structure, fund it properly, and leave it alone for two weeks before you judge it.

FAQ

How many campaigns should a new ecommerce Google Ads account have?

Start with three: Performance Max for your best-margin products, a Standard Shopping campaign with price-tier ad groups for control and data, and a small Branded Search campaign. Three campaigns is enough structure to get clean data without spreading a limited budget too thin. A $50/day budget split across eight campaigns gives each one so little traffic that none of them exit the learning phase properly.

What budget split should I use across these campaigns?

For a typical first-month budget, I run roughly 50% to Performance Max, 40% to Standard Shopping, and 10% to Branded Search. If your branded search volume is naturally high because you have existing brand recognition from social or PR, cap that campaign's spend rather than letting it balloon, since branded clicks are usually incremental sales you'd have gotten anyway.

Should I use Target ROAS from day one?

No. Run Maximize Conversion Value with no target until you hit 30-50 conversions per campaign. Setting a Target ROAS before the algorithm has real data usually causes it to under-spend and under-deliver, because it's trying to hit a number it has no evidence for. Once you have real conversion data, set your target based on your actual breakeven ROAS, which you calculate from your margin, not a number you picked because a competitor mentioned it on a podcast.

How do I handle a product feed with 500+ SKUs if I can't manually isolate all of them?

You don't need to isolate all of them. Isolate your top 10 by expected revenue and let the rest sit in catchall product groups segmented by price tier and category. Revisit monthly and promote any SKU that's quietly performing well into its own isolated group. This is exactly how I handled a catalog with hundreds of SKUs on a brand I scaled past $3M, and the 80/20 rule held up every single month.

The takeaway: build your google ads account structure around price tier and margin, isolate your real winners, keep branded spend small and separate, and give the account two full weeks before you touch it again. Structure is not the exciting part of Google Ads, but it's the part that determines whether your data means anything six months from now.