Same Google Ads Account, Two “Truths”: What Changes When You Audit for Sign‑In vs Purchase

Introduction: Why the conversion lens changes the entire audit

This case study comes from an anonymized, growth-focused advertiser that looked “busy” inside Google Ads—lots of conversion actions, mixed campaigns (web + app), and enough activity to create a sense that optimization was happening.

But when we ran two audits on the same account using different evaluation lenses—Sign‑in (a softer action) vs Purchase (a hard outcome)—the account told two different stories.

For local businesses, this matters even more than for e-commerce. When your real goal is calls, form fills, bookings, or paid appointments, letting Google optimize toward “easy” actions (page views, first visits, app sign-ins) is one of the fastest ways to inflate your CPA and still feel like you’re “getting conversions.”

The account context (generalized)

Think of this as a local-services business with both web and app touchpoints (common in healthcare, home services, local subscription businesses, etc.). There are multiple “conversion” events firing—some meaningful, many not.

The team had CPA targets set, but performance didn’t behave like a stable lead generation engine:

• In the Sign‑in lens (Dec 2025), the selected-action CPA was well above target (1,038 vs 250) even while the system recorded 238 conversions.
• In the Purchase lens (Aug 2026), the selected-action CPA was also above target (2,020 vs 1,500) on 161 conversions, with clear signs of waste.

The key detail isn’t the exact CPA—it’s the pattern: a lot of “conversions” can coexist with a broken optimization loop.

Lens #1: Audit the account as “Sign‑in” (December)

In the Sign‑in lens, the audit surfaced a paradox: the account was recording plenty of conversion activity, but structurally it was not in a state you’d consider “optimizable.”

Symptoms seen through the Sign‑in lens

• The audit flagged no active campaigns, meaning there was no live delivery to evaluate.
• At the same time, the conversion setup showed a large volume of primary conversion actions.
• The result was a misleading sense that the account had “conversion performance,” but little evidence that Google Ads was learning from the right outcomes.

Diagnosis process: why a Sign‑in lens can mislead

When we evaluate toward Sign‑in, we’re already one step removed from revenue. That’s not automatically wrong—some businesses do need sign-ins.

But the real issue here wasn’t “Sign‑in vs Purchase” yet. It was that the measurement layer was set up in a way that made nearly any user activity look like success.

Root causes found (Sign‑in lens)

1. Primary conversions were polluted with low-intent events

- Page view events like page_view and first_visit were set as Primary conversions.

- That tells Smart Bidding: “These are wins. Go get more of these.”

2. Too many primary conversions (31 active; recommended max ~3)

- Smart Bidding doesn’t “choose the best one.” It blends signals.

- With dozens of primaries, the system tends to chase the easiest-to-generate events (often the least valuable).

3. Attribution & measurement was effectively broken

- The Attribution/Measurement module scored 0 with a high volume of issues.

- Even if campaigns were active, the account would struggle to optimize toward a clean outcome.

4. Conversion counting was incorrect for a lead/signup behavior

- A sign‑in action was set to Every conversion.

- For sign-ups/leads, this commonly overcounts repeat actions and inflates “conversion volume,” distorting CPA and bidding.

In this lens, the account wasn’t “underperforming.” It was not measurable enough to trust the performance story at all.

Lens #2: Audit the account as “Purchase” (August)

In the Purchase lens, the audit got more “real.” Campaigns existed, there was sufficient data confidence, and we could see which segments were creating waste.

Symptoms seen through the Purchase lens

• The audit detected significant monthly waste.
• CPA was above target with meaningful conversion volume, but performance was uneven:

• Some campaigns were relatively efficient (best CPAs in branded/search-like segments).
• Others were dramatically inefficient (worst CPA campaigns several multiples higher).

This is a common local-business pattern: branded demand looks decent, while non-brand or broad intent campaigns quietly burn budget.

Root causes found (Purchase lens)

1. Attribution pipeline problems (app measurement blind spot)

- The audit flagged that a mobile measurement partner’s postbacks to Google Ads appeared inactive.

- Translation: Google Ads may not have been receiving the conversion signals it needed to bid properly for app outcomes.

2. Too many primary conversions (9 active; still above recommended)

- Even though this was “better” than 31, it’s still enough to dilute optimization.

- In purchase mode, this often causes Google to optimize toward proxy events (like sign‑in) because they’re cheaper and more frequent.

3. Smart Bidding deployed without enough goal-aligned volume

- Multiple campaigns used Target CPA or Maximize Conversions with single-digit purchase conversions for the month.

- Smart Bidding generally needs stable, real outcome volume (often ~30+/month per campaign as a rough rule of thumb) to behave predictably.

4. Keyword-level waste: spend above target CPA with zero goal-aligned conversions

- The audit identified keywords spending beyond the threshold while producing no purchase-aligned conversions.

- This is one of the cleanest “stop the bleeding” signals in a local account.

In the Purchase lens, the account wasn’t just “tracking wrong.” It was optimizing wrong—using advanced bidding on shaky data, while leaky targeting kept spending without producing the true outcome.

Side-by-side: what changed between lenses (and why it matters)

Here’s the practical comparison that local business owners and marketers can use.

1) Business outcome definition changes what “waste” looks like

• Sign‑in lens: waste shows up as measurement chaos and conversion inflation.
• Purchase lens: waste shows up as real money spent without goal-aligned results (keywords/campaigns exceeding CPA thresholds).

2) Different modules become the “main problem”

In Sign‑in mode, the biggest blockers were:

• Conversion Tracking Integrity (page_view/first_visit as Primary; too many primaries)
• Campaign Structure / Bidding (no active campaigns to evaluate)
• Attribution & Measurement (0 score; many issues)

In Purchase mode, the biggest blockers shifted to:

• Bidding & Smart Bidding Health (0 score; many critical issues)
• Attribution & Measurement (postbacks inactive; still 0 score)
• Audience & Targeting Quality (keywords spending above target with zero goal-aligned conversions)

3) CPA “truth” changes when counting is wrong

If you count page views and first visits as primary conversions, CPA can look artificially low or confusingly high, depending on how the “conversion” denominator is behaving.

Once you evaluate toward Purchase, you often discover the uncomfortable truth:

• Some campaigns were never designed to produce purchases.
• Some campaigns could produce purchases but are bidding/targeting with bad feedback signals.

What we would fix first (practical plan for a local business)

You don’t fix everything at once. You fix the feedback loop first, then you fix the levers.

Step 1: Clean up Primary vs Secondary conversions

• Make Primary conversions only the 1–3 outcomes that directly map to revenue.
• For local businesses, that’s usually:
• Booked appointment / completed lead form
• Qualified phone call (if tracked properly)
• Purchase/payment (if applicable)
• Move micro-events to Secondary:
• page_view
• first_visit
• time on site
• sign‑in (unless sign‑in is the business outcome)

If Google’s definition of “success” is wrong, every optimization that follows is directionally wrong.

Step 2: Fix conversion counting rules

• For lead/signup actions, set Counting to One per click (or one per interaction) to prevent inflation.
• Keep true purchases as Every conversion.

Step 3: Repair measurement / attribution plumbing (especially app-to-ads)

If you use an external measurement partner or offline conversion imports:

1. Confirm partner linking is active.
2. Confirm postbacks/imports are firing for the correct events.
3. Verify that the conversions appear inside Google Ads and are included in “Conversions.”

Step 4: Simplify bidding until volume is real

If a campaign only has a handful of purchase conversions per month:

• Avoid forcing Target CPA on sparse data.
• Consider stabilizing with:
• Manual CPC + enhanced signals (in some cases)
• Maximize Clicks (short-term testing)
• Broader portfolio approaches only after measurement is clean

Step 5: Stop the bleeding in targeting

• Pull a list of keywords/ad groups that:
• spent above your CPA threshold
• produced zero goal-aligned conversions
• Then take one of three actions:
• tighten match types
• add negatives based on search terms
• pause terms that repeatedly fail

Lessons local businesses can steal from this case

• A “conversion” is not automatically a business outcome. If page_view is Primary, you’re paying for traffic and calling it leads.
• Smart Bidding is only as smart as your conversion setup. Bad signals create expensive learning.
• Two audits of the same account can be right—because they answer different questions.
• Sign‑in lens asks: “Can we drive activation?”
• Purchase lens asks: “Can we drive revenue?”

Pick the lens that matches what you need to pay rent. Then make Google optimize for that.

If you want to sanity-check your own account

If you’re a local business running Google Ads and you’re seeing “conversions” without consistent revenue, the fastest win is usually a conversion goal cleanup and a bidding reset—before you touch creative.

If you want, you can try the VEOtool beta audit flow (discussion-first, no hard sell) or request an anonymized audit review to spot:

• which conversions are polluting Primary
• where Smart Bidding is running without enough signal
• where spend is drifting above your CPA threshold with no real outcomes

Soft CTA: If you’d like, share what your true business outcome is (calls, forms, bookings, purchases) and what Google Ads is currently counting as a “conversion”—I’ll tell you which lens you’re accidentally optimizing for.