PPC Advertising
Automated bidding is only as good as the data you feed it.
Paid search and shopping, managed against margin rather than platform-reported ROAS.
310k
Non-converting spend identified
Identified at account audit [PLACEHOLDER: reporting currency to be defined]
The problem
Smart Bidding took the manual work out of paid search and moved the leverage somewhere less visible. The algorithm optimises faithfully towards whatever you tell it a conversion is worth. Most accounts tell it something wrong: every lead weighted identically regardless of whether it closed, refunds and returns never subtracted, brand and non-brand blended into one target, Performance Max quietly absorbing branded searches and reporting them as new demand. The account looks healthy. The auction is being won at prices nobody chose.
Process
How we approach it
Fix the signal before the spend
The opening phase is conversion tracking, not campaign changes. We audit every conversion action, what it is worth, whether the value is dynamic, whether tax is included in it, and whether returns and cancellations are subtracted. Where sales close after the click, we import the outcome back from your CRM or order system so bidding optimises to revenue that survived, not to form submissions.
Separate defended demand from new demand
Brand search is usually the cheapest and least incremental spend in the account. It flatters blended ROAS and hides the performance of everything else. We split brand and non-brand reporting permanently, then test what actually happens when brand spend is reduced, rather than assuming either way.
Structure for control, not for tidiness
Consolidation is the platform’s preference, not always yours. We decide campaign structure based on where you need separate budget control and separate bid targets — then use exclusions, feed segmentation and negative lists to stop automated campaign types from eating traffic that a cheaper campaign would have won.
Review on a schedule, not when something looks wrong
Bid strategies drift. Search terms mutate. Feeds break silently. The work is a scheduled review with a documented reason for every change, so a performance shift can be traced back to a decision rather than guessed at three months later.
Deliverables
What you get
A full account audit at the start of the engagement, covering conversion actions, value assignment, bid strategy targets, attribution settings and asset group structure
Conversion value mapping so Smart Bidding optimises to margin or qualified-lead value rather than raw conversion count, with consumption tax stripped out where your store passes gross values
Offline conversion import configured from your CRM or order system into Google Ads and Microsoft Ads
Bid-strategy review on an agreed cadence: target CPA/ROAS adjustments, portfolio changes, and a written reason for each
Scheduled search-term review with negative keyword list maintenance at account and campaign level
Shopping feed diagnostics and title, attribute and disapproval remediation for retail accounts
Brand and non-brand performance reported separately as standard, never blended into a single ROAS figure
Industries
Where we run this
Ecommerce
Paid acquisition managed against contribution margin and breakeven efficiency, not platform-reported return.
SaaS
Demand capture and creation for B2B software, measured on payback duration rather than lead volume.
Finance
Acquisition for regulated firms, where the advertisement is itself a regulated document.
Gaming
User acquisition for free-to-play titles, measured on retention curves and payback windows.
B2B
Paid acquisition for considered purchases with long cycles, small keyword universes and offline conversions.
Frequently asked
We already run Performance Max. Will you keep it?
Usually yes, but not in its current shape. Performance Max works when it is fenced — brand exclusions applied, feed segmented, asset groups split by margin band rather than by product category. Left open, it absorbs branded and remarketing traffic and reports it as prospecting. We keep the campaign type and take back the controls.
How quickly will we see a change?
Tracking and structural fixes come first, and optimisation follows them rather than running alongside. Bid strategies then need a learning period after any material change, so the honest answer is that the first clean set of comparable data does not arrive in the first period of work, and the period you should judge us on is later still. We would rather say that at the start than let you draw conclusions from a fortnight of relearning.
Start with the audit.
It has a defined scope and a defined deliverable, and it is deliberately separable from anything that follows. If the audit says your current setup is fine, that is a legitimate outcome and we will say so.