Analytics
GA4 campaign benchmarks and Ask Advisor
GA4 campaign benchmarks only help if the tracking underneath is clean. I am a Gold Coast web developer with 18 years of storefront work. I have seen merchandisers treat a red "below peer group" tile as a brief to rebuild the site, when the real issue was a missing purchase event or a property that still mixed app and web. In 2026 GA4 will happily compare you to a modelled industry. It will not tell you that your industry dropdown is wrong.
Google's GA4 help for the Analytics product is the starting point for how the interface thinks. Benchmarking and Ask Advisor sit on top of the same property you already misconfigured. This note is how I set industry, how I read a comparison without panicking, and when I ignore the tile until GA4 ecommerce events match the gateway.
Ask Advisor is useful as a query box over your data. It is dangerous as an authority. It will summarise what GA4 believes. If GA4 believes you have no checkout, Advisor will invent a story about checkout.
What GA4 campaign benchmarks actually compare
GA4 campaign benchmarks actually compare your property to a peer group Google constructs from other properties that opted in and that Google has classified as similar. You are not seeing a named competitor. You are seeing a distribution. Traffic source, engagement and conversion metrics can appear as "your value versus typical range" depending on what Google exposes in that release of the product.
That comparison is only as good as three settings.
Industry category: if you sell sleepwear and the property is still tagged as general shopping or as a media site from a migration, the peer group is fiction. I set industry from how the business actually makes money, not from how the holding company is registered.
Business size and other qualifier fields: when GA4 asks, answer honestly. A three-person brand compared to enterprise marketplaces will always look "below" on volume metrics that are not the point.
Data quality: if half your sessions are internal, paid spam, or unfiltered staging hits, you will look like a different company. If conversions are mis-fired, you will look like a genius or a disaster. Benchmarks will not say "your tag is wrong." They will say "purchase rate."
Campaign-level reading needs campaign names that exist. UTM discipline still matters. If paid Search lands without utm_id or auto-tagging, GA4 will dump spend into (direct) or (not set). You then benchmark a channel that is really a tagging hole. Google Ads linking should be on, with auto-tagging, so that campaign, ad group and creative can populate.
I never use benchmarks to set a target CPA for Google Ads. Ads has its own auction. GA4 has a different attribution window and a different conversion definition. Reconcile, then decide. See GA4 attribution models for ecommerce.
On retail conversion at scale we tracked sessions, conversion, funnels and cart abandonment as operational numbers against the Magento stores, not against an anonymous peer cloud. Year-on-year sales moved $2.5 million over three years because the catalogue and the journey moved. A benchmark tile would not have told us which category to merchandising-test next. The funnel would.
When GA4 campaign benchmarks mislead a merchandiser
GA4 campaign benchmarks mislead a merchandiser when they mix incompatible journeys. A Shopify store with Shop Pay express will show a shorter funnel than a Magento site with a six-field checkout. Peer "begin checkout to purchase" rates are not a moral score. They are a mixture of vertical, device, and whether wallets exist.
They mislead when the property includes a blog subdomain, a wholesale portal, and the retail shop. Engagement time looks academic. Conversion rate looks tiny. Split properties or use data filters and reporting identities you understand.
They mislead when seasonality is ignored. A peak week compared to a trailing peer average will look like a miracle. A trough week will look like a crisis. Read year-on-year for the same store first. Then glance at peers.
They mislead when Ask Advisor is asked "why did conversion drop" after a CMP launch. Advisor may talk about landing pages. Tag Assistant would have talked about consent. Humans still have to pick the tool.
They mislead when enhanced measurement is counting scroll and outbound clicks as if they were store intent. A content-heavy Magento CMS can look highly engaged and sell nothing. A tight Shopify PDP can look cold and convert. Know the template.
I also see people screenshot a benchmark, send it to a developer, and ask for "more conversion." That is not a ticket. A ticket is: purchase revenue is 40 percent below gateway, or add_to_cart died on mobile Chrome, or paid landing URLs have INP of 400ms. Benchmarks can start a conversation. They cannot finish a deploy.
How Ask Advisor fits next to a human reading of campaigns
Ask Advisor fits as a faster way to pull a breakdown you could have built in explorations: top campaigns by purchase revenue this month, landing pages with high sessions and low add-to-cart, device split for paid. I treat the answer as a draft SQL brain, not as an audit.
I always open the underlying report. If Advisor cites a metric you do not collect (a key event you never marked), ignore the paragraph. If it cites a date range you did not mean, ignore the paragraph.
For campaign work I still export Google Ads and GA4 side by side. Advisor does not know your margin. It does not know that Performance Max is allowed to assist. It does not know that brand Search should not be cut because last click looks expensive.
Prompt like an engineer. "Show purchase key events by session campaign for the last 28 days, web only, excluding internal." If the product cannot filter that, you learned the limit. Do not keep rephrasing until it hallucinates a number.
I disable or ignore Advisor in client playbacks if the room will treat the sentence as fact. I will use it in my own prep to find anomalies, then I verify.
Production checklist before you trust a peer comparison
A property is ready for GA4 campaign benchmarks when the following are true.
- Industry category matches the live store.
- Google Ads linked, auto-tagging on, unwanted referrals filtered (payment gateways).
- Internal traffic defined. Staging blocked or in a separate property.
- Key events:
purchaseis the money event. Newsletter is not equal to purchase. - Ecommerce item data present so product-assisted questions have meaning.
- Currency and timezone match finance.
- Consent Mode not silently dropping all analytics for a region you sell into.
- At least a few weeks of stable tagging so you are not benchmarking a migration week.
- Channel grouping understood. Default channel group is not your media mix if UTMs are chaos.
- A written note of what you will ignore (volume vs enterprise peers, app+web mashups).
If you changed the measurement ID last month, do not trend across the change without a caveat.
What breaks the comparison overnight
A new GTM container that double-fires key events. You leap above peers. Someone asks what genius campaign did that. It was a duplicate tag.
A CMP default denied with no update. You fall below peers. Someone asks for a homepage redesign.
A Shopify hydrogen or headless cutover that loses the dataLayer. All ecommerce events vanish. Benchmarks show a collapse. The store still sells.
Renaming campaigns every week. GA4 cannot trend "Campaign 12" into "PMax - apparel." Your history fragments. Peers look stable because they are not you.
Marking every click as a key event. You outperform on "conversions" and underperform on revenue. Merchandisers celebrate. Warehouse does not.
How to measure your store first, peers second
Measure gateway revenue, then GA4 purchase, then Ads conversion value, then peer tiles. I wrote the same order into Magento retail operations and I still use it on contract work.
Look at campaign benchmarks as a smell test. If you are far outside the band on engagement but revenue matches the gateway, your vertical's peer group may include content sites. If you are far outside on conversion and revenue does not match the gateway, fix tags. If both match and you are still below peers, then look at landing pages, offer and feed. That is real work.
Do not publish peer percentiles as a case study number. I only publish lifts that exist on the work cases: Their Nibs storefront outcomes, and retail conversion sales. Benchmarks are a reading aid.
Related work on this site
Retail conversion at scale is the Magento operations case where funnels and conversion were managed against real sales, not against an anonymous cohort. Use that mindset when GA4 campaign benchmarks flash red. For the event schema underneath, read GA4 ecommerce events that match real revenue. For how credit is split across paid and organic, read GA4 attribution models for ecommerce stores.
FAQ
Are GA4 campaign benchmarks a ranking of my Google Ads account?
No. GA4 campaign benchmarks compare Analytics metrics in a peer group. Google Ads has its own auction insights. A campaign can look weak in GA4 last click and strong in Ads data-driven value. Read both.
Should I change the site because GA4 campaign benchmarks are below average?
Not until purchase revenue matches the gateway and the industry category is correct. Most "below average" tickets I see are tagging or consent. After that, look at landing pages and offer.
Can Ask Advisor replace an analyst for GA4 campaign benchmarks?
No. Ask Advisor can draft a breakdown. A human still has to know that Shop Pay skips a funnel step and that (not set) is a UTM problem.
How do I set industry so GA4 campaign benchmarks are less wrong?
Set it to the closest vertical the store actually is, in the Admin property settings Google provides, and revisit after a pivot. A fashion brand that added a journal should still be retail, not publishing, unless the journal is the business.
Keep reading
Contact if you want this kind of work on a live store.