Learn B2B SaaS Marketing

How to Diagnose B2B Paid Media Attribution Gaps

by 

Areej Gutale

Last updated: September 8th, 2026

Updated August 2026

Your paid media dashboard says one thing. Your CRM says another. Sales says the leads are junk. Finance says the pipeline is thinner than the reports show. Somewhere between the ad platform and closed revenue, signal is going missing, and every team is defending a different version of the truth.

That gap has a name. In our diagnostic framework it is called Loop, and it is the fifth and most under-diagnosed dimension of a paid-search program that has stopped scaling.

This guide walks the diagnosis. It starts with the symptom, maps it to the framework, sorts you into a tier based on how complex your CRM stack is (not how much you spend), walks the five canonical Loop questions, and gives you a fix sequence you can run yourself before hiring anyone.

Symptom to STALL map: why attribution is a Loop problem

Powered by Search runs paid programs through a five-dimension diagnostic called STALL: Sight, Traction, Acceleration, Lane, Loop. It finds the constraint that caps growth in a paid-search system. Each dimension asks one question, produces one KPI, and prescribes at most three sequenced actions.

Here is the full frame:

Attribution problems live almost entirely in the fifth row. When Loop fails, everything downstream distorts. Smart Bidding optimises for the wrong signal. Cost per opportunity climbs. Sales complains the leads are junk because bidding is pursuing a proxy the CRM never validated. Marketing cannot defend budget because the numbers on the dashboard are not the numbers in the pipeline. Finance sees fees rise while pipeline stays flat.

The other four dimensions can mask a Loop problem, which is why most teams misdiagnose. A Sight problem (impression share loss) looks like an attribution gap because visibility is dropping. A Lane problem (spend routed to brand or unproven demand) looks like an attribution gap because expensive clicks are not producing pipeline. An Acceleration problem (a landing page counting downloads as conversions) is often the direct cause of the Loop problem, because the wrong commercial action is what the platform learned to chase.

So the discipline is: work Loop first. If Loop is clean and pipeline is still flat, then work the other four. Do it in the other order and you will spend three months rewriting landing pages while the platform quietly bids toward a signal that has no relationship to revenue.

Rule of thumb: if your paid channel reports look healthy and your pipeline reports do not, the gap is Loop until proven otherwise.

Channel-boundary caveat

STALL was built as a canonical diagnostic for Google paid search. The current rubric applies there directly, with thresholds, benchmarks, and a validated payoff model. The Loop discipline itself, the five canonical questions, and the signal-maturity concept apply universally across paid channels. LinkedIn Ads, Meta, and Microsoft Advertising all support offline conversion imports, all bid against the signal you send them, and all fail in the same ways when the signal is wrong.

When you apply Loop to LinkedIn or Meta, the diagnostic thinking transfers; the specific benchmarks and platform mechanics do not. This guide uses Google paid search terminology throughout because that is where the rubric was proven. Where a step needs a channel adaptation, we flag it inline.

Which tier fits you? Sort by CRM complexity, not spend

Attribution failure modes track CRM stack complexity, not media budget. A $10K/mo advertiser on Salesforce with three custom objects has a harder Loop problem than a $150K/mo advertiser on HubSpot Starter with one pipeline. The Loop signal has to travel through your CRM before it reaches the ad platform, and every join, custom object, or funnel branch is a place the signal can drop.

Pick the tier that describes your CRM, not your spend:

Every tier has the same five Loop questions to answer. The likely gap differs by tier. The fix sequence differs by tier. The point at which it makes sense to bring in an outside team differs by tier.

The five canonical Loop questions

Every Loop diagnosis walks these five questions in order. Each has a specific data source, a specific pass or fail criterion, and a specific fix if it fails.

1. Is the primary goal a real commercial action? A demo booked, a trial started, a sales conversation held. Not a form fill, not a PDF download, not a pricing-page view. If the primary goal is a proxy, Smart Bidding is optimising to buy more proxies, and every downstream metric distorts from that root.
2. Are there conflicting goals bidding is optimising against? When two goals both count as primary at different points in the account (one at the account level, another via CustomConversionGoal at the campaign level), bidding pursues an inconsistent target. Signal noise climbs, cost per opportunity climbs, and no fix outside the goal structure will hold.
3. Is the offline conversion import fresh? Anything older than seven days is stale for a system that bids in real time. A configured integration is not a working integration. Freshness is the test.
4. Are conversion values reflecting real pipeline economics? If every conversion is imported at a value of 1, bidding treats a demo with a $500K deal size the same as a demo with a $5K deal size. Value-based bidding needs values. Values should reflect qualified pipeline, not raw form fills.
5. Is bidding actually using the signal? Smart Bidding strategies (Target CPA, Target ROAS, Maximize Conversions, Maximize Conversion Value) must be pointed at the goal that carries the real signal. If the signal is imported but the campaign is on Manual CPC, or on Smart Bidding pointed at the wrong goal, the platform learns nothing.

Every tier walks these five questions. The tier tells you where the failure usually lives.

Simple tier: the proxy problem

You are on HubSpot Starter or Pro with a single pipeline. You have one sales team. Your ads platform is Google, LinkedIn, or both. You are spending anywhere from a few thousand a month to well into six figures. The dashboards report conversions steadily. The sales team is unimpressed by the leads.

Walk the five questions.

Question 1: Is the primary goal a real commercial action? In the Simple tier this is where the diagnosis ends more than half the time. The primary conversion action is a form fill (contact form, gated PDF, "download the guide"). Smart Bidding optimises for volume of that form fill. Sales cannot close a form fill. Marketing cannot report a form fill as pipeline. The signal loop closes on a proxy that has no economic meaning downstream.

Fix: define one commercial action. In HubSpot the cleanest anchor is Meeting Booked from a specific meeting link tied to the sales team, or a Lifecycle Stage change to Sales Qualified Lead. In your ad platform, configure that action as the primary conversion via the HubSpot-to-Google or HubSpot-to-LinkedIn native integration. Downgrade every proxy action (form fill, download, page view) to Secondary or Observation so bidding stops chasing them.

Question 2: Conflicting goals? In the Simple tier, unlikely. You have one pipeline, one sales team, one funnel. Verify by opening Google Ads → Goals → Conversions and confirming only one action is marked Primary at the account level.

Question 3: Import freshness? Native HubSpot-to-Google and HubSpot-to-LinkedIn integrations sync on schedules measured in hours, not days. Confirm the last successful sync in HubSpot → Marketing → Ads → Sync status. If it has failed, the fix is a five-minute re-auth.

Question 4: Conversion values? In the Simple tier most advertisers do not use value-based bidding at all, and that is fine at this stage. Volume of one clean commercial action beats value-based bidding on a noisy signal. Do not layer value complexity onto a Loop that has not yet been fixed at the goal level.

Question 5: Is bidding using the signal? Once the primary goal is a real commercial action, switch bidding to Maximize Conversions with a Target CPA equal to the current average, or Target CPA at the level that keeps volume steady. Give it three weeks to relearn. Do not touch it during learning.

KPA for Simple tier, in order:

1. Redefine the primary conversion action as one real commercial action (meeting booked or SQL stage change).
2. Downgrade proxy actions to Secondary or Observation.
3. Switch bidding to Smart Bidding pointed at the new primary goal; hold the account for three weeks.

Expected effect: cost per commercial action becomes a defensible number by week four. Sales-quality complaints track down within one full sales cycle. Pipeline reporting stabilises because the metric on the dashboard is the metric in the CRM.

When to bring in an agency at Simple tier: you probably do not need to. This is a self-serve fix. If you are spending under $15K/mo, run the three steps yourself and revisit in ninety days. If pipeline is still flat after a clean commercial-action goal, that is the point to talk to someone.

Medium tier: the freshness and CustomConversionGoal problem

You are on HubSpot Pro or Enterprise, or Salesforce with one custom object. You have two or more sales teams, or two or more funnels feeding one pipeline. You have already configured offline conversion imports. Someone on your team believed the tracking was working.

Walk the five questions.

Question 1: Primary goal. In the Medium tier the primary goal is usually a real commercial action (SQL, Opportunity Created, Meeting Booked). If it is not, walk back to Simple tier first. If it is, keep going.

Question 2: Conflicting goals. This is where Medium-tier Loop problems concentrate. CustomConversionGoal at the campaign level lets you point one campaign at a different conversion action than the account primary. When teams inherit accounts, CustomConversionGoal overrides accumulate. Different campaigns end up optimising against different definitions of success. Signal fragments. Bidding never fully converges.

Diagnosis: in Google Ads, open every top-spend campaign and check the Goals section for a CustomConversionGoal override. In our audit process, we walk the top-spend campaigns in the UI, campaign by campaign, because API data alone cannot prove signal absence. A campaign can report the account-primary goal in the API while the UI shows a CustomConversionGoal override that is what bidding is actually using. This is our proprietary discipline. Competitors reading API dumps miss the gap.

Fix: align every top-spend campaign to the account primary. Remove CustomConversionGoal overrides unless you have a specific, defensible reason (a distinct sales team with a distinct pipeline that genuinely needs separate optimisation).

Question 3: Import freshness. The Medium-tier gap. A configured import is not a working import. The Salesforce-to-Google Click ID (GCLID) match rate degrades over months as sales-side users update records outside the flow that carries the GCLID. The HubSpot-to-Google integration silently drops rows when field mapping breaks after a HubSpot schema change.

Diagnosis: Google Ads → Tools → Conversions → click any offline conversion → "See recent uploads." If the latest upload is older than seven days, or if the volume of imported conversions dropped significantly week over week without a matching drop in pipeline, freshness has failed.

Fix: re-auth the integration. Confirm the GCLID field is populated on the Lead or Opportunity object at the moment sales marks it qualified. If the GCLID is going missing between form submission and qualification, that is a workflow bug in the CRM, not a Google Ads bug.

Question 4: Conversion values. In the Medium tier, values are usually all set to 1, and that is often acceptable if your deal sizes cluster tightly. Layer value complexity only after freshness and goal alignment are clean.

Question 5: Bidding using the signal. Confirm every top-spend campaign is on Smart Bidding, pointed at the aligned primary goal. If a campaign is on Manual CPC because someone wanted to protect it during a learning-phase reset, that campaign is not benefiting from the fixed Loop.

KPA for Medium tier, in order:

1. Align top-spend campaigns to the account primary goal in the UI; remove unjustified CustomConversionGoal overrides.
2. Re-auth the CRM-to-platform integration and confirm import freshness under seven days.
3. Move any Manual CPC top-spend campaigns to Smart Bidding pointed at the aligned goal.

Expected effect: cost per opportunity stabilises within one full learning cycle (typically three weeks). Sales-marketing arguments about lead quality quiet down because both sides are looking at the same signal. The pipeline number on the ad dashboard and the pipeline number in the CRM converge.

When to bring in an agency at Medium tier: this is where an outside diagnostic pays off. The CustomConversionGoal audit is tedious. The freshness diagnosis touches CRM workflows that marketing does not always own. A STALL audit closes the gap in weeks rather than quarters and hands you a fix sequence your team can execute. If you are spending $20K/mo or more, this is the tier where a paid audit earns its cost back inside a quarter.

Complex tier: the goal-conflict and value-misconfiguration problem

You are on Salesforce with multiple custom objects. You may have CPQ. You may have product-led signals mixed with sales-led signals. Multiple pipelines feed one revenue number. Different business units bid on different keywords toward different commercial actions.

Walk the five questions.

Question 1: Primary goal. In the Complex tier the primary goal debate is not a technical question, it is a business question. Which commercial action is the one the platform should learn to pursue? If you are running product-led motions (trial started) alongside sales-led motions (demo booked), you have to decide which one carries the load for Smart Bidding, or split the account so each motion has its own campaigns and its own aligned goal. Trying to bid against both from the same campaign structure is where Complex-tier Loops break.

Fix: decide the goal at the business level, then reflect it in the account structure. Product-led campaigns bid against trial started, valued at pipeline-equivalent. Sales-led campaigns bid against demo booked, valued at pipeline-equivalent. Do not blend them.

Question 2: Conflicting goals. In the Complex tier this is the compound version of the Medium-tier problem. Multiple CustomConversionGoal overrides across business units, sometimes inherited from a re-org, sometimes left in from a legacy campaign. UI-level audit is the only way to see the full picture. We open every top-spend campaign, note the goal in force, and produce a conflict map.

Question 3: Import freshness. In the Complex tier freshness is a shared responsibility problem. Salesforce workflows touch the GCLID field, Opportunity object, and Lead object across teams. When any one of those workflows changes, freshness can silently degrade. Diagnostic discipline: monthly freshness audit, held by the marketing operations owner, tied to a named person in Salesforce admin.

Fix: monthly review of upload success rate and GCLID match rate. Set an alert on match rate below 60%. When it fires, treat it as a P1 incident.

Question 4: Conversion values. This is the Complex-tier fix that unlocks the most upside. If your deal sizes range from $10K to $500K, bidding on volume of demos is buying the wrong demos. Value-based bidding, fed by real deal-size data from Salesforce, lets Smart Bidding pursue the demos that resemble your $500K deals rather than your $10K deals.

Fix: map deal size or expected deal size to the imported conversion value. Feed the real number, not a proxy. Switch bidding to Maximize Conversion Value with a Target ROAS calibrated to your close rate and gross margin. This is where the Complex tier stops looking like "we have attribution" and starts looking like "we have attribution that teaches the platform to buy revenue, not clicks."

Question 5: Bidding using the signal. In the Complex tier the failure mode is old bidding strategies persisting on top-spend campaigns because no one wanted to touch them. Confirm every top-spend campaign is on the correct value-based bidding strategy pointed at the correct goal.

KPA for Complex tier, in order:

1. Decide the primary goal per business motion (product-led vs sales-led) and reflect it in campaign structure.
2. Audit every top-spend campaign in the UI for CustomConversionGoal conflicts; align to the correct motion primary; enforce monthly freshness review.
3. Move to value-based bidding fed by real deal-size data from Salesforce; calibrate Target ROAS to close rate and gross margin.

Expected effect: the pipeline mix shifts toward larger deals within two full sales cycles. Cost per opportunity may rise. Pipeline value per dollar of spend rises faster. Finance can defend the paid budget against the real revenue number, not against a proxy.

When to bring in an agency at Complex tier: this tier is the reason integrated pods exist. The Loop diagnosis touches Salesforce administration, sales operations, marketing operations, and paid media in the same week. Trying to hold all four contexts inside one internal team is where most Complex-tier programs stall. A STALL audit produces the fix sequence; a delivery pod runs it across the four teams without dropping context between them.

Why UI verification is proprietary discipline

The single most under-appreciated point in the diagnostic is this: API data alone cannot prove signal absence. A CustomConversionGoal override at the campaign level is visible in the Google Ads UI. It is inconsistently exposed in the API depending on which endpoint you query. Auditors who pull an API dump and grep for the account primary conversion will report a false clean bill of health.

We walk top-spend campaigns in the UI, campaign by campaign. Every campaign gets its goal setting read from the screen, not inferred from a report. This is slow. It is also the only way to catch the CustomConversionGoal blind spot before it costs you a quarter of misdirected bidding.

Any audit that promises to diagnose Loop from an API pull alone is telling you something about the audit, not about your account.

The other four dimensions, briefly

Loop is where attribution problems live, but the other four dimensions can mask a Loop problem or be its underlying cause. Walk them briefly once Loop is clean:

• Sight. If non-brand impression share lost to Ad Rank is above 30%, the auction is capping visibility before Loop can teach the algorithm anything. Fix Sight (Quality Score, bids, eligibility) alongside Loop.
• Traction. If non-brand high-intent CTR is below the vertical benchmark, the ad is losing to competitors on messaging or SERP coverage. Fix creative and ad extensions.
• Acceleration. If the landing page CVR to the one commercial action is below vertical benchmark, the Loop is fixed but the funnel narrows before conversions can flow. Fix landing pages second.
• Lane. If more than 30% of non-brand spend is going to brand-harvesting keywords or unproven demand-creation terms, the budget is not routed to demand that can become pipeline now. Rebalance allocation.

The rule stays: work Loop first, work the other four second. Doing it in the other order is how programs spend six months rewriting landing pages while the platform quietly bids toward a proxy.

The verification checklist

Before you conclude your Loop is clean:

• One primary conversion action per campaign, and it is a real commercial action (demo, trial, sales conversation), not a proxy.
• No CustomConversionGoal overrides on top-spend campaigns unless there is a specific business reason, verified in the UI, not the API.
• Offline conversion import last successful sync is under seven days.
• Conversion values reflect real pipeline economics (deal size or expected deal size), not all set to 1, if you are using value-based bidding.
• Every top-spend campaign is on Smart Bidding pointed at the aligned primary goal.
• The pipeline number on the ad dashboard and the pipeline number in the CRM agree within a small margin.

If you cannot check every box, you have a Loop problem. The tier you fit tells you where to start.

When to bring in Powered by Search

• Simple tier, under $15K/mo: run the three-step KPA above yourself. Revisit in ninety days. If pipeline is still flat after Loop is clean, that is the point to have a conversation.
• Medium tier, $20K/mo and up: a STALL audit is the right paid entry point. We walk the UI, produce the conflict map, hand you a sequenced fix your team can execute, and price the audit to earn its cost back inside a quarter. See how Powered by Search runs a STALL audit for lean B2B teams.
• Complex tier: the audit is the diagnostic; the delivery pod is the execution. A lean B2B pod pairs a Director of Demand Generation with a paid search and social performance marketer, a design lead, and a development lead. No account director, no project manager, no junior bench. Direct-to-SME delivery, because Salesforce administration, sales operations, marketing operations, and paid media have to move in the same week to close a Complex-tier Loop, and every layer of coordination is a week you do not have.

The Powered by Search delivery model is a lean B2B marketing team wanting a strategic, proactive, self-driving partner with the widest integrated scope under one umbrella, bought as pipeline as an outcome. Attribution is one of the four disciplines the pod holds. Google Ads, LinkedIn, landing pages, and attribution move together, not as four vendors on four workstreams.

Related comparisons

If the diagnosis surfaces a data or CRM problem alongside the paid gap, these hubs cover the adjacent selections:

• Best B2B data analytics agencies. When the Loop failure traces to broken reporting infrastructure rather than paid configuration
• Best B2B HubSpot agencies. When the freshness or workflow gap lives inside HubSpot
• Best B2B Salesforce agencies. When the Complex-tier fix needs a Salesforce administration partner alongside the paid pod
• Best B2B marketing automation agencies. When the Loop failure is downstream of a marketing automation misfire

Attribution is not a reporting problem. It is a signal-routing problem, and the signal has to travel from your ad platform to your CRM and back before Smart Bidding can pursue anything real. Diagnose Loop first. Fix it in the sequence your tier calls for. Bring in help at the point where the diagnosis outruns what an internal team can hold at once.