The short answer
Improving ROAS without increasing budget comes down to spending the same money on better clicks, with better creative, on better landing pages, with better bidding. Most UK accounts have 25-40% wasted spend hiding in plain sight - search terms that don't convert, dayparts that lose money, geographies that produce junk leads, broken landing pages. Find and fix those before asking for more budget.
Below are the 12 highest-leverage optimisations Ben uses on UK accounts, ranked roughly by typical impact. None require additional spend.
What ROAS actually measures (and what it doesn't)
ROAS (Return on Ad Spend) = Conversion Value / Ad Spend. A ROAS of 4.0x means every £1 of ad spend produced £4 of revenue. It's the most useful single ecommerce metric because it ties spend directly to revenue.
What ROAS doesn't tell you:
- Profit - a 4x ROAS on a 20% margin product is breakeven; on a 60% margin product it's gold
- Incrementality - high ROAS on brand campaigns often reflects users who would have bought anyway
- Lifetime value - a 2x ROAS on a subscription product with 12-month average tenure is excellent
For lead-gen accounts, replace "Conversion Value" with "Pipeline Value" or "Closed-Won Revenue" via offline conversion import. Otherwise you're optimising towards form fills, not revenue.
Diagnose before you optimise
Before changing anything, run these four queries against the last 90 days:
- Search terms by spend, no conversions - sort search terms report by Cost descending, filter Conversions = 0. This is your wasted-spend list.
- Hour-of-day performance - segment by hour. Look for hours where Conv Rate is < 50% of average.
- Geography report - segment by region/city. Look for areas with >5% of spend and <1% of conversions.
- Device report - segment by device. Look at the gap between mobile and desktop conversion rates.
Within 30 minutes you'll have a prioritised list of where the money is leaking. Fix the largest leaks first.
1. Aggressive negative keyword strategy (typical impact: 10-20% wasted spend recovery)
The single highest-leverage optimisation in most UK accounts. Build three negative keyword lists:
- Universal negatives - applied account-wide. Examples: "free", "cheap", "DIY", "tutorial", "jobs", "salary", "wikipedia", "reddit", competitor brand names
- Category-specific negatives - applied to specific campaign groups. E.g. for a B2B SaaS: "students", "personal", "home use"
- Search-term negatives - added weekly from the search terms report
Use phrase or exact match negatives - broad negatives can accidentally block valuable queries. Review the search terms report at least weekly. Tools like Opteo and Optmyzr automate the surfacing step (covered in Ben's separate Optmyzr vs Opteo post).
2. Weekly search terms audit (typical impact: compounds with negatives)
If you only do one weekly task in Google Ads, make it the search terms audit. Filter the search terms report to the last 7 days, sort by Cost descending, then for each row decide:
- Is this a converting term? → Add as exact match keyword in a relevant ad group
- Is this irrelevant? → Add as a negative
- Is this borderline? → Leave it for now and re-evaluate in 14 days with more data
This single discipline keeps an account clean over time. Most accounts that "stop performing" after 6 months stopped because nobody did the search terms audit.
3. Move to tROAS bidding (typical impact: 15-25% conversion value uplift)
If you have 100+ conversions/month per campaign and accurate conversion values flowing in, switch from Maximise Conversion Value (no target) to tROAS. Set the target at your trailing 60-day average ROAS minus 10% (to give room for growth).
tROAS forces Google to defend your efficiency floor. Without a target, Smart Bidding will happily chase incremental conversions at any ROAS. Read more in Ben's separate post on when to trust the Google Ads algorithm.
4. Pass dynamic conversion values (typical impact: enables tROAS, 20%+ value uplift)
If your Google Ads conversion sends a flat value (e.g. every lead = £100), you're throwing away signal. Real businesses have a wide range of conversion values - a £5,000 enterprise lead vs a £200 SMB lead should not be valued the same.
Pass dynamic values via the data layer (ecommerce) or via lead scoring (lead gen). Google's value-based bidding setup guide covers the implementation. Once dynamic values flow, tROAS becomes an order of magnitude more effective.
5. Dayparting and ad scheduling (typical impact: 5-15% efficiency gain)
Most accounts have predictable performance patterns by hour of day and day of week. Examples Ben sees in UK accounts:
- B2B SaaS: weekday 9am-5pm converts 3x better than weekend evenings
- Local services: 7am-9am Monday traffic converts at half the rate of midday traffic (commuters browsing)
- Ecommerce: 8pm-11pm browsing peaks but converts lower than morning shoppers
Use ad schedule bid adjustments (-30% on low-performing dayparts, +10-20% on high-performing ones). On Smart Bidding, dayparting still works but as observed signals - Smart Bidding factors them in automatically. Pause campaigns entirely during dayparts that lose money consistently.
6. Geographic bid adjustments (typical impact: 5-15%)
Same diagnosis as dayparting but for location. Pull the geographic report, find regions with <0.5x average conversion rate, apply -50% bid adjustment or exclude entirely. Find regions with >1.5x average conversion rate, apply +20% bid adjustment.
UK accounts often find concentration in London + Manchester + Birmingham. Setting +20% bid adjustments in those cities while -30% in low performers can lift overall ROAS 5-10% with no other changes.
7. Audience signals and exclusions (typical impact: 5-15%)
Add audience signals to PMax campaigns. These don't restrict targeting but tell Google "users like these convert" - speeding up the learning algorithm. Best signals:
- Customer Match list of past purchasers
- Website visitors who hit high-intent pages (pricing, demo, checkout)
- In-market segments matching your category
Equally important: exclude audiences that won't convert. Add existing customers to acquisition campaigns as exclusions to stop wasting spend on people who already bought. Add competitor employees if you can identify them via Customer Match.
8. Landing page conversion rate uplift (typical impact: 15-40%)
The most under-attended ROAS lever. A 30% conversion rate uplift on the landing page directly translates to 30% better ROAS at the same spend. Common UK landing page wins:
- Mobile speed - if Largest Contentful Paint > 2.5s on mobile, fix that first. Use PageSpeed Insights.
- Form length - cut every non-essential field. Each removed field typically lifts conversion 5-10%.
- Trust signals above the fold - reviews, certifications, "as seen on" logos.
- Clear CTA - one primary action, repeated 2-3 times down the page.
- Mobile design - 70% of Google Ads traffic in the UK is mobile. Build mobile-first, then desktop.
9. Ad assets - sitelinks, callouts, structured snippets (typical impact: 5-10% CTR)
Ad assets (formerly extensions) are free. Every account should have:
- 4-6 sitelinks per campaign
- 4-8 callouts
- 1-2 structured snippets
- Lead form (where relevant)
- Call asset (where relevant)
- Image assets (PMax / Display)
Higher CTR means higher Quality Score means lower CPC. The ladder works. Don't have any assets? You're paying 10-30% more per click than you need to.
10. RSA optimisation (typical impact: 5-15%)
Responsive Search Ads use 15 headlines and 4 descriptions, mixed and matched by Google. Optimisations Ben uses:
- Pin one branded headline to position 1 (controls perception)
- Test promotional headlines vs benefit-led headlines in 30-day cycles
- Aim for "Excellent" Ad Strength in every ad group
- Use the asset performance report to identify "Low" performers and replace them monthly
11. Shopping feed optimisation (typical impact: 10-30% for ecommerce)
Shopping campaigns are 80% feed quality, 20% bidding. Wins:
- Title rewrites with brand + product type + key attribute (size, colour, material)
- Detailed product descriptions matching common search terms
- High-quality product images, white background, multiple angles
- Custom labels for margin/ROAS-based campaign segmentation
- Fix all Merchant Center warnings
Tools like DataFeedWatch and Channable automate feed optimisation at scale.
12. Offline conversion import (typical impact: 30-50% real ROAS uplift)
For lead-gen accounts the single biggest ROAS lever isn't optimisation - it's swapping the conversion you optimise towards. Stop optimising to form fills, start optimising to qualified leads or closed deals via OCI. Ben's separate post on how to track offline conversions covers the full setup.
Which to do first
If you only have a week, do these in order:
- Search terms audit + universal negatives (1 hour, 10-20% recovery)
- Geographic and dayparting cleanup (1 hour, 5-15%)
- Mobile landing page audit (3 hours, 15-40%)
- Ad assets - fill every gap (1 hour, 5-10%)
That's 6 hours of work, no budget increase, typical combined uplift 30-60% ROAS.