.png%3F2026-11-17T10%253A32%253A11.150Z&w=3840&q=75)
It’s easy to be sceptical of Google Ads these days. In the last 5 years alone, Google have pushed advertisers onto broad match, rolled out PMAX without decent reporting in place and launched AI Max, but LinkedIn is already full of critics and un-actionable articles so we’re going to aim to give you real insights into Google Ads performance and what new products to use this Black Friday.
AI Max For Shopping
If you’re like us and want to prepare your clients to be able to meet the growing demand in AI-driven search results this Black Friday, read on...
Before launch, give AI guardrails to the freedom it’s allowed with your brand:
Set a ROAS target from day one
Term exclusions for competitors, retailers and brands, so you’re not matching to searches for stockists or rival products
Messaging restrictions in text guidelines (still in beta), telling AI Max to:
URL exclusions for broad pages such as: contact, blog, login and about pages, so final URL expansion only sends people to pages built for purchase
A brand exclusion for your own brand list, with Shopping ads on excluded brands left switched off, so you’re measuring new, non-brand demand rather than brand searches you’d likely win anyway
Review daily SQRs for the first two weeks, adding exclusions as soon as anything irrelevant creeps in
Our results
- Spend up 540%, unlocking the headroom Shopping had never had next to PMax
- Revenue up 285%, driven by highly relevant long-tail AI search queries we weren’t reaching before
- ROAS down 40%, but comfortably within target, and a significant improvement on the initial two week learning period
tROAS vs Maximise Conversion Value
We wanted to understand if since the 17th August bidding update, whether removing ROAS constraints during a sale would let two hero product campaigns capture more of the extra demand. In one sale, both campaigns ran on target ROAS. In the other sale, we switched both to Maximise Conversion Value with no ROAS target, so Google could optimise to drive as much revenue as possible.
Our results
Hero product one
- Spend up 301%
- Revenue up 16%
- ROAS down 71%
Hero product two
- Spend up 229%
- Revenue up 119%
- ROAS down 33%
Across both campaigns, spend was up 3.6x and revenue up 70%, with Google ROAS down 53% on maximise conversion value compared to tROAS. So Maximise Conversion Value did exactly what it says on the tin: it spent the budget chasing every bit of revenue it could, with no regard for efficiency.
So during a sale, when auctions are more competitive and CPCs are already climbing, it ends up being a costly way to grow volume therefore we recommend keeping target ROAS switched on through peak. If you want to push for more volume, use Smart Bidding Exploration (see below) lower the target slightly for the sale window rather than removing it entirely, and use a seasonality adjustment if you’re expecting a sharp jump in conversion rate over a short period.
Smart Bidding Exploration (and Promotion Mode)
In the same sale where Maximise Conversion Value hit our hero product campaigns so hard, we tested Smart Bidding Exploration for the first time across another two campaigns.
Rather than removing the ROAS target entirely, Smart Bidding Exploration lets Google flex it to explore new searches it would otherwise have held back on. For the purpose of our test, we set that flex to the maximum 30%.
Our results
Campaign one, period on period:
- Spend up 32%
- Revenue up 29%
- ROAS down 2%
Campaign one, compared to the previous sale:
- Spend down 28%
- Revenue down 12%
- ROAS up 22%
Campaign two, period on period:
- Spend up 21%
- Revenue up 22%
- ROAS up 1%
Campaign two, compared to the previous sale:
- Spend up 10%
- Revenue up 26%
- ROAS up 14%
Where Maximise Conversion Value pushed spend up 3.6x for a 70% lift in revenue, Smart Bidding Exploration grew spend and revenue in line, so ROAS held across the sale. The only dip was campaign one spend and revenue against the previous sale, but even there the revenue we did drive came in far more cost effectively.
So if you want to scale through peak, we still don’t recommend you throw away your ROAS target to do it. Instead we recommend you use Smart Bidding Exploration to find new searches while still bidding to a target, so you get volume without sacrificing efficiency.
What this all means for Black Friday
If you want Google Ads working hard over peak, you need to be launching these final test and learns now. Black Friday is on 27th November this year, and with a two week learning period you want that well behind you before demand spikes. Set the guardrails we’ve suggested above, and we hope you’ll have a successful, and stress-free (as much as possible) peak period.
Ready to grow?
Want results like this?
Book a free discovery call and we'll show you exactly how we'd approach your growth.
More to read
Related articles

Cold outreach isn't dead, lazy outreach is
by Gaby Rollinson
.png?2026-09-23T15%3A18%3A29.165Z)
Most ecommerce brands will reset their Demand Gen learning before peak. Here's what Google says to do instead
by Chloe Lodge

What does HubSpot's State of Marketing 2026 report tell us about high intent buyers and AI search?
by Gaby Rollinson
Let's build something that performs
Whether you're scaling up or shaking things up, we'd love to chat about how we can help. No hard sell — just a conversation about your goals.
