Before raising the budget, name where the extra spend will go. More of the same high-intent demand is a different proposition from broader searches or another market. The account average cannot price demand you have not bought yet.
The next €2,000 needs its own business case
Use these hypothetical mature outcomes to separate average performance from the value of an increment.
| Period | Spend | Qualified demos | Cost per qualified demo |
|---|---|---|---|
| Baseline | €10,000 | 50 | €200 |
| Expanded | €12,000 | 55 | About€218 |
| Increment | €2,000 | 5 additional | €400 |
The blended result looks only slightly worse, but the additional demos cost €400 each. That does not automatically mean stop. Compare the increment with your acceptable acquisition economics, opportunity quality and available sales capacity. It does mean that the baseline€200 figure is a weak forecast for the next budget increase.
I would first identify where the extra spend went: the same intent group, broader searches, another market or different delivery. Separate brand from non-brand where that distinction matters, and reconcile qualified-demo counts with the CRM. Do not count an imported outcome twice because it also appears in a lead report.
This before-and-after example illustrates arithmetic, not causal proof. Demand, competition and sales behaviour may have changed. A bounded experiment with a suitable baseline gives a stronger basis for the decision.
Accept a higher demo cost when better opportunities justify it. Do not use the account average to excuse weaker demand.
Earn the next budget increase
Average efficiency answers what the account has achieved. Marginal efficiency asks what the next investment bought. The two can move in different directions while the headline report still looks healthy. Keep the added spend and qualified outcomes visible, then examine their quality and timing. A higher marginal cost can be acceptable for better opportunities. It should be an explicit commercial choice, not a surprise hidden inside the blended cost per conversion.
Check what the platform calls a conversion
List the actions included in bidding and reporting. Are duplicate events counted? Are lightweight actions treated like valuable enquiries? Does the event mean what the business thinks it means?
A campaign can efficiently produce the wrong event. Fixing measurement is not cosmetic reporting work; it changes what the account is rewarded for finding. Keep a record of tracking changes so apparent performance shifts are not mistaken for market changes.
Separate demand with different economics
Review brand and non-brand performance, important intent groups and the actual searches you can inspect. Strong existing brand demand can make a blended report look reassuring while broader acquisition remains weak.
Search visibility is incomplete in some reporting contexts, so use the available evidence without claiming you can inspect every query. Look for repeated patterns of irrelevant or commercially weak demand.
Check the page and the sales handoff
Does the landing page continue the ad promise? Can the appropriate buyer understand the offer and act on mobile? Does the lead reach the right owner quickly enough?
If you already lose good-fit visitors after the click, additional traffic may increase the waste. If sales cannot respond to the current volume, the next budget increase needs an operational plan too.
Review landing-page friction and sales-accepted lead quality together. A budget increase should not amplify a broken handoff or a page that attracts the wrong expectation.
A scaling decision sheet
| Check | Evidence to gather | A reason to pause |
|---|---|---|
| Measurement | Validated event definitions | Duplicates or misleading primary actions |
| Demand | Intent-level performance | Expansion into weak-fit searches |
| Quality | Mature CRM outcomes | More leads without useful conversations |
| Conversion | Page and device performance | Broken or confusing next steps |
| Capacity | Follow-up and delivery readiness | Good demand cannot be handled |
No single row gives automatic permission to scale. Review the whole picture and identify the likely constraint.
Test the next increment
Choose a bounded increase or expansion with a clear hypothesis. State which demand you expect to reach and what would count as useful progress. Review marginal outcomes, not only the old account average.
If additional spend produces fewer qualified opportunities per euro, inspect where it went. It may reflect a temporary learning period, changing competition, broader intent or a real limit in the current approach. Do not immediately call every decline “learning.”
Make an imported demo pass a record-level check
Choose an authorised test record with a known campaign journey and qualification outcome. Trace the identifier, actual outcome time, imported event and reported count. Confirm that the same event is not represented twice in the measurement used for the decision.
Check unsuccessful form attempts as well as successful submissions. A thank-you page that can be reopened should not create a fresh sales result every time someone visits it. The implementation needs a dependable success signal.
Only after those checks should you compare quality by intent group or adjust the optimisation target. An elegant bidding plan trained on inconsistent events is still an inconsistent plan. Keep a short validation record so the next account change does not quietly undo the measurement work.
Check the current measurement settings before following an old tutorial
Google’s enhanced-conversions guidance describes a unified setting for web and leads, with migration beginning in April 2026. Older setup screenshots may therefore differ from the current account interface.
Check the account’s actual configuration and current documentation before changing the implementation. Then validate the business outcome separately: the imported event must still represent the intended qualified action, with appropriate data handling and deduplication. A newer measurement feature does not repair an inconsistent sales definition.
Audit the conversion setup as a business system
Review the actions included in reporting and bidding, their definitions and how they are counted. Follow a real or authorised test journey through the page and relevant systems. Confirm that a successful action is recorded once as intended and that failed actions do not masquerade as success.
Check changes in the conversion setup against the performance timeline. A new event, altered counting rule or changed primary action can move the reported cost per conversion without changing the business result. The team needs to know what changed before interpreting the graph.
Keep a simple measurement record with event name, meaning, owner and validation method. This helps future account managers understand the setup rather than inheriting a list of actions whose original purpose has been forgotten.
Examine the economics of the next demand group
Use mature quality outcomes where available. Compare the groups by useful conversations, opportunities and the later commercial information the business can reliably identify. Include deal differences when they materially affect value.
| Expansion question | Evidence to collect | Decision it supports |
|---|---|---|
| Can existing demand absorb more spend? | Delivery constraints and quality | Increase within the current scope |
| Is a new intent group relevant? | Query context and sales feedback | Run a bounded test |
| Does the page serve the new audience? | Message fit and mobile journey | Adapt before expanding |
| Can sales handle the volume? | Ownership and response capacity | Coordinate operations |
| Is the signal reliable? | Validated events and timing | Keep or revise optimisation |
Use the table to organise a decision, not produce a mechanical score. Some evidence will be incomplete. State the uncertainty and scale the test to the risk.
Prepare the landing page for the new promise
A new audience may need a different explanation even when the product is unchanged. Review the problem language, examples, proof and action. A page that works for a familiar buyer may leave a newly targeted role unsure why the offer matters.
Check the mobile journey in particular. A broad campaign expansion can change the device mix and make an existing form problem more consequential. More traffic should not be sent into a page whose important action is difficult to use.
Use genuine evidence relevant to the buyer. Do not add impressive but unrelated results simply because the new page looks sparse. The proof should resolve a concern, not distract from the missing explanation.
Run a bounded pilot and keep the baseline visible
Define the scope, budget boundary, primary outcome and guardrails. Keep the original successful group identifiable so the pilot does not become impossible to evaluate inside an account-wide average. Record material changes and avoid changing several important variables at once unless the work genuinely requires it.
If the pilot underperforms, review whether the hypothesis failed or the implementation was incomplete. A broken form invalidates a conversion comparison. A weak audience may invalidate the expansion premise. Different failures require different responses.
Do not rescue every disappointing test by increasing the budget. More data is useful only when the test remains relevant and the potential learning justifies the cost. Establish a reason to continue, modify or stop.
Why the bidding switch came later
In the Google Ads quality rebuild, qualified-demo measurement came before a controlled bidding pilot. The case describes why limited, delayed deep-funnel events needed care: a more meaningful for the business signal still had to be dependable enough to use. That distinction matters when implementing the diagnostic check above. A desirable business outcome and a ready optimisation event are two different things.
Before approving the increase, write the stop condition
A budget proposal should name the search theme receiving more money, the outcome it is expected to add, and the condition that would reverse the decision. “Scale what works” leaves too much room for interpretation.
Use a short decision record: baseline period, expansion period, spend, mature qualified demos, sales acceptance, known changes in demand and the person responsible for reviewing the result. Choose the review window around the actual sales delay. A Friday report cannot judge enquiries that sales has not yet qualified.
| Proposed move | Evidence needed before release | Reason to pause |
|---|---|---|
| Increase an existing high-intent campaign | Reliable outcomes and sales capacity | Incremental demos become too expensive |
| Add broader searches | Query review and a separate budget | Repeated weak fit despite relevant wording |
| Change the bidding outcome | Validated imports and usable event volume | Missing, duplicated or delayed signals |
The qualified-demo case illustrates the sequencing. Measurement and search intent were repaired before deeper-funnel bidding was piloted. Lower qualified-demo cost came with flat media spend; it was not evidence that simply increasing the budget would have produced the same improvement.
I would rather approve a smaller increase with a clear rollback rule than a large increase defended by an attractive account average. The average describes money already spent. The proposal concerns money the business has not spent yet.
The better-qualified signal can still be the wrong next switch
A qualified-demo event can be meaningful for the business but too sparse, delayed or inconsistently recorded for a dependable change in optimisation. Fixing the definition does not finish the implementation.
Validate record-level imports, timing and counting. Keep the intermediate signal visible while testing whether the deeper outcome can support the intended decision. The qualified-demo case illustrates why measurement repair preceded a controlled pilot.
If quality improves under the pilot without undermining delivery, review the case for expansion. If the data cannot support a confident judgement, revise the test rather than defending a full rollout. The aim is better opportunity creation under workable constraints, not the deepest possible event merely because it sounds closest to revenue.
A budget increase needs delivery capacity
Qualified demand can improve while the sales team becomes unable to respond quickly enough. Before expanding a campaign, review the number of enquiries the team can process, the response route and the next appointment available. An increase that creates a queue may reduce the quality of the buyer experience and obscure the campaign’s contribution. Keep the scaling decision with both the media evidence and the operating constraint. A bounded increase, agreed review point and visible backlog are more useful than treating the platform recommendation as a complete commercial plan.
Sources and further reading
Google Ads: offline conversion imports: Connecting later outcomes with ad measurement.
Google Ads: enhanced conversions for leads: Lead measurement and implementation guidance.
References checked 9 October 2026. Sources support the documented facts and tool behaviour; the recommendations and worked examples explain how I would apply them.
Budget increment decision
Use the blank sheet to organise your own evidence and decisions. The column headings follow the diagnostic method in this guide.
Download the working sheet ↓CSV format. Opens in Excel, Google Sheets and other spreadsheet tools. Name, email, company and location required before downloading.
Calculate the cost of the next increment
The starting figures are worked examples. Replace them with your own data. Values are calculated in this page and are not sent anywhere.
Marginal cost = additional spend ÷ additional qualified demos. Both increments must be positive. Use the same currency throughout.
Was a step unclear or missing? Tell Amine what you were trying to do ↗. Your email can include the step, what you expected and what prevented you from using it.
Has the next budget increase earned its place?
Before raising spend, bring the campaign split, current conversion definitions and the sales outcome you want more of. We can assess what the next budget increment would need to prove.