Back to insights

    Performance marketing · 11 min

    Google Ads leads but few buyers? Measure quality through sales feedback

    Measure the journey from a received enquiry to a qualified opportunity and won business before deciding which advertising signal to optimize.

    Dark spheres beside three metal filters and a small group of emerald glass spheres on a raised platform.

    Cheap enquiries can still be expensive to sell to

    When Google Ads produces form submissions but few suitable buyers, start by checking what the account rewards. A submitted form proves that someone made contact. It does not prove that your offer fits their needs, that a real buying project exists or that sales has reviewed the enquiry. Measure lead quality through a shared qualification definition, a reliable link to sales outcomes and feedback into advertising.

    This approach is particularly useful for B2B services, complex projects and sales involving several steps between enquiry and order. The purpose is to identify campaigns that create opportunities your team can actually serve. If enquiries remain unanswered, the first improvement may belong in the sales process. More advertising will not resolve that operational gap.

    Before adjusting budgets, connect three observations: what the visitor wanted, what they submitted and what happened afterwards. Without the third, campaigns can become better at finding cheap contacts while sales spends more time on unsuitable conversations. The workflow below is NUMEDIA's editorial working framework, not a forecast of performance improvement.

    Give every stage a meaning and an owner

    Start with a short agreement between marketing and sales. A “good lead” is not a useful data category until two people can classify the same enquiry consistently. Write down the need you serve, the customer types and locations you support, and the evidence required for a meaningful next step. Define qualification around the offer, not around making a report look successful.

    Google distinguishes leads qualified outside its advertising platform from converted leads that reach a later milestone. That milestone does not automatically mean a paid order: the business must define it. In your own reporting, keep accepted opportunities, issued proposals and won deals separate, even when only one stage is used for campaign optimization.

    A workable internal sequence is received enquiry, valid enquiry, qualified enquiry and accepted opportunity. A valid enquiry is neither spam nor a duplicate. Qualification confirms a relevant need and fit with the offer. An accepted opportunity has an agreed next step. Record the date, responsible person and reason whenever the stage changes so that acquisition and handling problems remain distinguishable.

    Allow an “outcome unknown” state. An unanswered call is not evidence that someone is an unsuitable buyer. The timing, message or review may be wrong. Report unresolved outcomes separately and assign responsibility for resolving them. Otherwise, campaign comparisons can end up measuring individual salespeople's follow-through more than the quality of incoming demand.

    Read five measures together

    Cost per lead is advertising spend divided by unique received enquiries from a defined source and period. Cost per qualified enquiry uses the same spend and qualified enquiries from that same group. Qualification rate shows the share meeting your agreed criteria. Alongside these, track the proportion not yet reviewed and the proportion that becomes won business.

    Make the denominator explicit. “Half our leads qualify” means something different when sales has assessed every enquiry than when it has assessed only the easiest people to reach. Show qualified enquiries as a share of all enquiries and as a share of reviewed enquiries, with the outstanding count beside them. A small sample does not become a reliable forecast simply because it is expressed as a percentage.

    Consider an illustrative calculation using invented figures. Campaign A spends EUR 1,000, receives 50 enquiries and produces 10 qualified opportunities: EUR 20 per enquiry and EUR 100 per qualified opportunity. Campaign B spends the same amount, receives 25 enquiries and also produces 10 qualified opportunities: EUR 40 and EUR 100 respectively. B is not automatically worse. This example excludes sales costs, contribution margin and completed orders.

    If both campaigns produce the same number of suitable opportunities, the workload required to assess them may matter. Measure that workload before claiming a saving. Likewise, a large proposal value is not revenue. Keep open pipeline and won deals distinct, with a clear observation date and consistent treatment of cancelled or reopened opportunities.

    A minimum record that connects the teams

    You do not necessarily need a large CRM project to start. You need a structured record that the team maintains. Google's import documentation allows lead capture systems other than a dedicated CRM. A spreadsheet can help validate definitions and ownership before an integration is commissioned, although access and data handling still need to be managed.

    Give each enquiry a stable internal ID. Record receipt time, requested service, available source, landing page, current stage, latest stage-change time, owner and disqualification reason. Do not put personal details into URLs or general analytics fields. Use a supported purpose-built connection for advertising measurement, with the identifiers and requirements applicable to the chosen method.

    Distinguish “source unknown” from “not from Google Ads”. A missing identifier is not permission to fill the gap with a salesperson's guess. Preserve the unknown category and measure how much of the pipeline can actually be connected. A rising matchable share can improve reporting coverage without demonstrating any improvement in advertising.

    Assign ownership at field level. Development owns capture and transmission, sales owns accurate stage changes, the advertising specialist owns conversion mapping, and the business owner approves the definitions. A technically healthy connection can still carry stale commercial information. Automating a poorly defined field only distributes its errors faster.

    Verify sales feedback before using it for bidding

    When sales confirms the agreed milestone, that event should reach the selected conversion action. Enhanced conversions for leads supplement offline event imports with matching information. They do not turn an unsuitable contact into a buyer. The event definition, capture process and maintained integration remain essential.

    Scope implementation against current documentation for the selected connection, such as Data Manager or a supported CRM integration. “Send all conversions” is not an adequate brief. Specify the event, actual event time, identity handling, duplicate rules, permitted data and recovery process for rejected transfers. Ordinary analytics parameters containing personal information are not a substitute for the supported implementation.

    Observe a new signal before allowing it to influence optimization. Google's implementation checklist recommends initial secondary-action validation and diagnostic checks. Your acceptance evidence should follow a specific record: it exists in the sales system, has the correct stage and time, is accepted by the import, and appears under the intended action. A green connection indicator alone does not prove commercial correctness.

    Test sending the same event again. An integration must recover from a failed job without presenting the same milestone as another opportunity. Test missing fields, changed stages and incorrect timestamps too. When reconciling records with Google Ads, use an appropriate conversion-time view and account for rejected or still-processing records rather than assuming every uploaded row was accepted.

    Choose the optimization event deliberately

    A primary action can be used for bidding when the campaign uses its associated goal. A secondary action is generally for observation. Custom goals are an important exception: an included secondary action can also be used for bidding. Check the settings of each affected campaign, not just the label next to an action in the account overview.

    Do not count a form submission, qualification and won deal as three independent customers. They may be three milestones on one journey. Reporting should show progression, while the optimization goal is selected deliberately. Check frequency, delay and reliability before changing it. The deepest event is not always the most usable signal if it arrives rarely or incompletely.

    For a long sales cycle, confirmed qualification may be a practical optimization milestone while won deals remain a later business measure. This is a process-specific decision, not a universal setting. There is no single conversion-count threshold appropriate to every account: campaign type, bidding approach, volume and data quality all affect what is workable.

    Document the date of a goal change and its expected reporting effect. Moving from submitted forms to qualified enquiries can reduce the displayed conversion count purely because the definition changed. That is not evidence of declining performance. Retain a comparison of the old and new events so management can understand the change without mistaking a measurement decision for a business trend.

    A decision matrix for the next intervention

    Review rejection reasons and handling status before changing ads. The following matrix is an original working template for a joint marketing and sales review. Support each diagnosis with records rather than an untested impression that all enquiries are poor.

    A shorter form will not solve slow sales responses. Faster responses will not solve a stream of enquiries for a service you do not offer. Address the narrowest confirmed constraint, then review the same measures again. Multiple simultaneous changes make interpretation harder, particularly when enquiry volume is small.

    NUMEDIA editorial decision matrix: evidence before intervention.
    ObservationEvidence to inspectFirst intervention
    Many enquiries for the wrong serviceEnquiry content and the advertising promiseAlign the offer, search intent and landing page.
    Many unreviewed enquiriesOwner and first-response timingFix handover and sales ownership.
    Many duplicatesRecord IDs and repeat submissionsRepair capture and counting rules.
    Qualifications missing from AdsMapping, event time and rejected importsRepair the connection before changing bidding.
    Suitable conversations, few winsLoss reasons and cohort maturityReview the offer, handling and sales cycle.

    Compare cohorts with equal time to mature

    Enquiries received this week have had less time to progress than enquiries received last month. Comparing won-deal rates without equal follow-up time penalizes newer campaigns. Group enquiries by receipt date and assess them at comparable ages. Choose an observation interval appropriate to the actual sales process and label it clearly.

    Separate the commercial record in the CRM from advertising attribution. Different dates, models and available identifiers can produce different totals. Reconciliation aims to explain the differences, not force two tables to agree. An attributed sale does not establish that the advertising caused an additional sale that would otherwise never have happened.

    A weekly operating review should establish whether records are arriving, whether sales is assessing them, which disqualification reason dominates and which single intervention will be checked next. A monthly review can add the cost of suitable opportunities and mature sales outcomes. At low volume, combine sensible periods and show counts alongside rates rather than presenting unstable percentages as precise evidence.

    What to commission, and when to wait

    A measurement brief should identify the current form, CRM or tracking record, sales stages and existing conversion actions. Do not send a personal-data export with an initial enquiry. A field description, anonymized journey example and system list are enough for an initial scoping discussion. Separate commercial definitions, technical connection, acceptance testing and ongoing monitoring in the requested scope.

    Ask the supplier for a mapping document, test evidence, named error ownership and maintenance instructions. Acceptance should also record which actions remain secondary, when a goal may change and who approves that change within your business. Asking only whether a tag has been installed is too narrow for a system expected to support better sales decisions.

    Full automation is premature if nobody will maintain outcomes or if the team disagrees about qualification. Establish the record and routine first. If reliable data already exists, campaigns produce a regular flow and manual reconciliation is becoming fragile, a dedicated measurement project can be the appropriate next step.

    For a discussion with NUMEDIA, start with your website address, a description of the sales process and the point where the link between advertising and business outcomes disappears. That provides a basis for a concrete review and implementation scope. The first objective is to identify the missing connection, not to promise a lower acquisition cost before the evidence exists.

    Sources and methodology

    1. About qualified leads and converted leads (Google Ads Help, accessed 2026-11-05)
    2. About conversion goals (Google Ads Help, accessed 2026-11-05)
    3. Enhanced conversions for leads implementation checklist (Google Ads Help, accessed 2026-11-05)
    4. Offline conversion imports FAQs (Google Ads Help, accessed 2026-11-05)

    Frequently asked questions

    Does a low cost per lead mean the campaign is working?

    Not on its own. Review qualification rate, cost per qualified enquiry, handling and later sales outcomes. Similar contact costs can conceal very different commercial value.

    Do we need a CRM before starting?

    Not necessarily. A structured record with stable IDs, stages, dates and owners can establish the process. Choose an integration when definitions are stable and manual work becomes a constraint.

    Should a sale immediately become the primary conversion?

    Check frequency, delay and data reliability first. Confirmed qualification may be more usable in a long cycle. Inspect campaign goals and the custom-goal exception as well.

    What should we do with enquiries whose source is unknown?

    Keep them in a separate unknown-source category. Do not guess that they came from advertising. Track matching coverage because changes in coverage affect report comparability.

    When can we judge a change?

    When compared cohorts have had enough time to move through the normal sales process and records are sufficiently complete. There is no universal deadline. Separate a new measurement definition from a business outcome.