PPC Advertising
Automated bidding is only as smart as the data behind it.
Paid search and shopping managed for margin, not platform-reported ROAS.
310k
Wasted spend uncovered
Found during account audit [PLACEHOLDER: reporting currency to be confirmed]
The challenge
Smart Bidding removed the manual effort from paid search and shifted the real leverage somewhere harder to see. The algorithm faithfully optimises towards whatever value you assign to a conversion. Most accounts feed it the wrong information: every lead valued the same whether it closed or not, refunds and returns never deducted, brand and non-brand merged into one target, and Performance Max quietly taking branded searches and reporting them as new demand. The account appears healthy while auctions are being won at prices nobody agreed to.
Our process
Our approach
Fix the signal before the spend
The first phase is conversion tracking, not campaign changes. We review every conversion action — its value, whether that value is dynamic, whether tax is included, and whether returns and cancellations are deducted. Where sales close after the click, we import results from your CRM or order system so bidding optimises for revenue that actually stuck, not form submissions.
Separate existing demand from new demand
Brand search is normally the cheapest and least incremental spend in an account. It inflates blended ROAS and masks how everything else performs. We permanently split brand and non-brand reporting, then test what really happens when brand spend is reduced, rather than assuming either outcome.
Structure for control, not neatness
Consolidation suits the platform, not always you. We design campaign structure around where you need separate budget control and separate bidding targets — then use exclusions, feed segmentation and negative lists to stop automated campaigns from taking traffic a cheaper campaign would have won.
Review on schedule, not only when something looks wrong
Bid strategies drift. Search terms change. Feeds break without warning. Our work follows a review schedule with a documented reason for every change, so any shift in performance can be traced to a decision rather than guessed at three months later.
Deliverables
What you receive
A full account audit at the outset, covering conversion actions, value assignment, bid strategy targets, attribution settings and asset group structure
Conversion value mapping so Smart Bidding optimises for margin or qualified-lead value rather than raw conversion count, with sales tax removed where your store sends gross values
Offline conversion imports set up from your CRM or order system into Google Ads and Microsoft Ads
Bid strategy reviews on an agreed schedule: target CPA/ROAS adjustments, portfolio changes and a written reason for each
Scheduled search-term reviews with negative keyword lists maintained at account and campaign level
Shopping feed diagnostics plus title, attribute and disapproval fixes for retail accounts
Brand and non-brand performance always reported separately, never merged into one ROAS figure
Industries
Sectors where we deliver this
Ecommerce
Paid acquisition managed to contribution margin and breakeven efficiency, not the return platforms report.
SaaS
Demand capture and creation for B2B software, measured on payback time rather than lead volume.
Finance
Acquisition for regulated firms, where every advert is itself a regulated document.
Gaming
User acquisition for free-to-play games, judged on retention curves and payback periods.
B2B
Paid acquisition for high-consideration purchases with long sales cycles, narrow keyword pools and offline conversions.
Common questions
We already use Performance Max. Would you keep it running?
In most cases yes, though not as it is set up today. Performance Max performs when it has guardrails — brand exclusions in place, the product feed segmented, and asset groups organised by margin tier instead of product category. Without them it soaks up branded and remarketing traffic and labels it as new-customer prospecting. We keep the campaign type and regain control of it.
How soon will we see results?
Tracking and structural fixes come first, with optimisation following rather than running in parallel. Bid strategies then need a learning period after any significant change, so honestly, the first clean set of comparable data will not appear in the opening period, and the period you should judge us on comes later still. We prefer to say that upfront than have you draw conclusions from two weeks of relearning.
Begin with an audit.
The audit has a clear scope and a clear output, and it stands on its own — no obligation to continue. If it shows your current setup is working well, that is a perfectly valid result and we will tell you so.