Google Ads Shopping Campaign Structure for Ecommerce

Sebastien Day

Apr 16, 2023 — 3 min read

I love ecommerce, and now work exclusively with ecommerce businesses after working at one for 18 years.

This account structure is one I use with my clients and own business. The idea is split the products into campaigns based on past performance, then use a Performance Max campaign to drive cheaper traffic for the poor performing products that are not getting clicks or sales.

All Products is a standard shopping that is a catchall (smaller budget, higher tROAS). This should not be a Pmax campaign, otherwise it will cannibalise the standard brand campaigns.

All Villains and All Zombies are both Pmax.

The rest of the campaigns, Brand 1 and 2, are Standard Shopping campaigns based on brand or product category. This allows heavy use of negative keywords and pushing a high ROAS for the key products.

The idea is that poorer performing products get some budget in the Pmax campaign, and then if they start to perform automatically move out of the Pmax and fire from the standard shopping campaigns.

The budgets will depend on the number of products in the campaign. I always give a campaign as much budget as it wants while it's hitting target ROAS.

Why not Pmax all the day down?

I get better performance with shopping campaigns.
You lose access to all that juicy search term data.
You can't add negatives to the campaign (just the account).
It's more resilient to changes to products, budgets & bids. You won't go back into learning from a simple budget increase.

Overview

Label your products in the merchant centre feed as a hero, sidekick, villain or zombie.
Setup a catch all Performance Max campaign that targets villains and zombies only.
Create Shopping campaigns on target ROAS based on brand or product category*

Conversion Tracking

Obviously an absolute must.

This seems to be a subject many in PPC don't fully understand as it's "technical". Learn about tag manager, data layers, events and GA4 and enough copy paste to implement them. You need to know how to implement and test tags.

You must track with values (the £££ amount)

add to baskets add_to_cart
start of checkout begin_checkout
ordered placed purchases

The best practice method is to use GA4 events that are then pushed to Ads as Conversions.

There are many other ways of tracking with custom JS, UA or gtag but if you use Google you need to get on the GA4 train.

Feed Structure

💡

Optimise the shopping feed before doing anything

The feed is the most important part of successful shopping ads - spend the time optimising the titles, descriptions and assigning to the correct google categories.

Label as Heroes & Villains

Segment your products into four groups of Heroes, Sidekicks, Villains and Zombies, that are what they say they are. Some Shopping CSS offer this as a service, or you can manually create a feed with the tags - it's just two fields, sku and custom_field_0.

Heros - perform above your target tROAS
Sidekicks - perform at or above 75% of tROAS
Villains - over x% of budget spent and not hero or sidekick
Zombies - everything else. They can't have had x% budget spent so low volume, and can't be converting otherwise would be sidekick as above 75% of tROAS

Performance Max

I consistently beat Pmax with a standard Shopping campaign, structured to break the products into logical groups with lots of negative keywords.

You also get a wealth of search term report data that can be used to optimise the products, which will then futher improve shopping and organic rankings.

That said Pmax works better as a "catch all" Shopping campaign for the products that aren't top sellers as I've found it can achieve a lower CPC and drive lots of traffic with a smaller budget.

Pmax algo and learning phases

This might shock you, but Google uses Performance Max as an excuse to rinse you for any spare change you may have. With a new campaign set a low budget and no target ROAS.

And expect Google to do some really stupid shit with your budget.

It will double spend for the first couple of days and use every penny available to test keywords that obviously won't convert. However it does level out after 5 days, you just have to sit it out.