How to Improve Google Ads Conversion Tracking Accuracy

If 29% of Google Ads accounts can go 90 days with zero conversions, the first question is not “how do we get more clicks?” — it is “are we even measuring the right conversions?” That number comes from WordStream’s 2026 Google Ads performance analysis, which reviewed more than 250,000 Google Ads Grader reports across 15,000+ accounts. For teams trying to improve google ads conversion tracking accuracy, that is the uncomfortable starting point: what looks like a bidding problem or a weak campaign often turns out to be a broken measurement system. We see this pattern constantly. Teams tweak audiences, rewrite ads, and debate attribution models while their primary conversion fires inconsistently, their CRM never closes the loop, and half their browser traffic disappears behind privacy protections.

The more useful way to frame the problem is this: tracking accuracy is not mainly a tagging task. It is a system design task. You need the right conversion event, a data path that survives modern browsers, and landing pages that make intent easy to capture and reconcile later. That matters even more now because Forrester’s 2024 analysis of cookie deprecation argues that signal loss will weaken Google’s ability to connect ad exposures to conversions, while remarketing data becomes less complete. In other words, the old “drop a tag and trust the platform” playbook is getting worse, not better.

Our view is straightforward and slightly contrarian: the best way to improve Google Ads conversion tracking accuracy is not to obsess over attribution models first. It is to make the conversion signal itself harder to break. That means redesigning the event, the page, and the data path around privacy loss, fraud risk, and messy user journeys. Once you do that, campaign optimisation becomes far more rational because you are no longer steering from distorted numbers.

Why your conversion data lies

WordStream’s 2026 performance study is blunt: the average business wastes $1,127.54 per month in Google Ads, and some accounts waste nearly half of their total search budget. The same dataset found that about 29% of accounts recorded zero conversions over 90 days. That is not just an efficiency issue. It is a measurement warning. When a channel with this much intent data reports nothing, the odds are high that the system is failing before the campaign is.

A lot of teams assume inaccurate conversion reporting shows up as obvious breakage. It usually does not. More often, data lies quietly. Form submits count twice. Qualified leads and low-intent leads both count as one conversion. Calls from ads appear in Google Ads, but never in the CRM. Offline deals close three weeks later and never get imported. The account looks “clean” because numbers exist. They are just the wrong numbers.

What does zero conversions actually mean?

Zero conversions does not always mean zero business impact. WordStream’s 2026 guide to Google Ads conversion tracking notes that conversion tracking exists to show how many clicks actually result in sales, but that only works if the conversion is defined correctly, tagged correctly, and attributed consistently. Their other 2026 performance piece also points out a simple explanation for many zero-conversion accounts: conversion tracking is not set up, so campaigns may have generated results the advertiser will never know came from Google Ads.

Consider a simple scenario. A SaaS company spends €8,000 per month on search. It gets 1,600 clicks at an average €5 CPC. Sales says the team booked 24 demos from paid search last month. Google Ads reports 3 conversions. The issue is not demand generation. The issue is that the reporting path from click to demo is broken.

A practical diagnostic looks like this:

• Google Ads clicks: 1,600
• Landing page sessions: 1,420
• Form starts: 110
• Form submissions in analytics: 31
• Lead records in CRM with paid-search source: 24
• Google Ads reported conversions: 3

In that chain, the campaign is clearly driving response. The break happens between form submission and ad platform attribution. Teams that stop at “Google Ads says 3” end up cutting useful campaigns.

The edge case matters too. Sometimes zero conversions really does mean zero conversions. If the account is sending broad informational traffic to a weak page, measurement is not the culprit. But you cannot know that until you compare click volume, on-page behaviour, and downstream lead creation together.

How much budget is bad tracking burning?

The waste is not limited to uncounted conversions. Bad tracking also creates false positives, which push more budget into low-quality traffic. If your account optimises toward newsletter signups, shallow page views, or duplicate form fires, Google’s bidding system learns the wrong lesson. It buys more of what is easy to count, not what creates pipeline.

Here is a worked example with real account math. Suppose you run €20,000 per month in spend and import one primary conversion called Lead. Google Ads reports 200 conversions, so your apparent cost per conversion is €100. Later, the CRM shows only 70 unique leads, because:

• 40 were duplicate submissions
• 55 were spam or junk entries
• 35 were low-intent ebook downloads counted as leads

Your real cost per unique lead is not €100. It is €20,000 / 70 = €285.71. If only 20 of those leads become sales-qualified, your true cost per SQL is €1,000.

This is why we treat tracking error as a hidden tax. It distorts bidding, reporting, forecasting, and landing page testing at the same time. It also infects adjacent decisions, including how you judge ad creative and whether you think your ad messaging is actually working.

There is a second hidden cost: fraud and invalid interactions. Harvard Business Review’s 2009 piece on online ad fraud reminds us that fraudulent or low-value ad activity is not a new issue. It has been a recognised problem for years. If your measurement stack is weak, you are not just missing conversions. You are also giving suspect interactions room to look legitimate.

That raises the next question. Before you rebuild tags, how do you tell whether the problem is tracking, traffic quality, or landing page friction?

Separate tracking errors from real performance

Most teams troubleshoot in the wrong order. They change bids first, then creative, then targeting, and only much later ask whether the conversion event fires consistently. We recommend the reverse. Start with a triage model that isolates measurement, traffic quality, and page friction before you touch budgets. Otherwise you optimise noise.

This is where our first operating framework helps.

The Track-Then-Trust Audit

The Track-Then-Trust Audit is a simple three-stage diagnostic. First, verify that the conversion event fires when the intended user action happens. Second, test whether behavioural signals on the page support the idea that traffic is qualified. Third, compare platform-reported conversions with downstream CRM outcomes before making any campaign decision. The point is discipline: do not trust a number until it survives all three checks.

Here is a practical scoring version you can copy this week:

Now apply it to a hypothetical B2B account:

• 900 ad clicks in 30 days
• 72 form starts
• 28 submitted forms recorded in analytics
• 14 Google Ads conversions
• 26 CRM leads with paid search source

That pattern tells you something specific. Tracking is undercounting. Traffic quality is probably not the primary issue because form starts and submissions exist at reasonable rates. The CRM count being closer to analytics than Google Ads points to a platform attribution or tag handoff problem.

Is it tracking or just bad traffic?

This is one of the most common search questions because the symptoms overlap. Flat conversion numbers can come from poor targeting, but they can also come from missing signals. WordStream’s 2026 benchmark report found the average Google Ads click-through rate is 6.66% and that 65% of industries saw better conversion rates in 2025. That matters because weak performance is not automatically evidence that Google Ads has become ineffective. In many sectors, conversion efficiency improved.

A clean way to diagnose the difference is to compare three ratios:

• Click to landing page session
• Session to form start
• Form start to form submit

If the first ratio collapses, tracking or landing page load issues may be blocking sessions. If the second is poor, the offer or intent match is weak. If the third is poor, the page experience or form UX is the likely culprit.

Consider this example:

• Clicks: 2,000
• Sessions: 1,150
• Form starts: 96
• Submits: 11

A 42.5% click-to-session loss is too large to ignore. Start with measurement and page delivery, not campaign bids. By contrast, if clicks and sessions align closely but form starts are near zero, the campaign may simply be attracting the wrong visitors.

The contrarian point is important here. Bad traffic is often blamed for what is really bad instrumentation. We have seen teams pause profitable keyword groups because mobile Safari underreported form events.

Which numbers should match and which never will?

This is another place where teams create unnecessary panic. Not every system should match exactly. Different tools use different attribution windows, identity rules, and bot filters. Expecting perfect parity between Google Ads, analytics, and your CRM is unrealistic.

What should match closely are counts of deterministic business events: confirmed form submissions, booked meetings, verified calls, and imported offline conversions tied to a click identifier or reliable first-party key. What will never match perfectly are sessions, assisted conversions, and view-through influenced actions.

A healthy operating tolerance looks like this:

• Form submits: within 5-10% across form backend, analytics, and server logs
• Google Ads primary conversions vs CRM leads: within 10-20% after attribution window adjustments
• Revenue figures: often wider gaps unless offline imports are disciplined

If your gap is 40-60%, do not treat it as normal attribution variance. That is a system issue.

Once you isolate the failure type, the next step is more fundamental than most guides admit: fix the conversion event itself.

Fix the conversion event itself

Most inaccurate accounts do not suffer from too few tags. They suffer from bad event design. WordStream’s 2026 conversion tracking guide outlines four main Google Ads conversion action types: website actions, phone calls, app installs and in-app conversions, and uploaded offline conversions. That is useful, but the more strategic question is which of those actions should count as success in your account and which should stay secondary.

A lot of teams count everything. Button clicks. Scroll depth. Whitepaper downloads. Chat opens. Calendar views. Then they ask why smart bidding chases low-quality users. The answer is obvious once you say it aloud: you taught the system that noise equals success.

Which actions should Google Ads count?

We recommend a simple hierarchy:

• Primary conversions: actions that strongly predict revenue and should guide bidding
• Secondary conversions: useful engagement signals for diagnostics, not bidding
• Noise events: interactions that should never be treated as business outcomes

HubSpot Community’s 2024 discussion on Google Ads conversion tracking points to a practical downstream view: track conversions using attribution reporting in ads, then mark contacts as customers when they become customers. That is the key distinction. A submitted form is not the same as a customer. If your setup cannot separate those stages, your reported performance will drift away from actual business value.

Here is what we actually use as a starting scoring model for lead-gen accounts:

If an account optimises to demo requests and qualified calls, bidding learns much faster than when all form fills count equally.

The edge case: PLG and ecommerce setups can justify lighter-funnel actions as primary conversions if they correlate tightly with purchases. For enterprise SaaS with six-month cycles, that shortcut usually creates expensive confusion.

How do you stop counting junk conversions?

Start by defining exclusion rules before you publish any tag. That means deciding what does not qualify. We prefer explicit filters such as:

• Exclude submissions with free email domains for sales-demo campaigns
• Exclude duplicate form submissions within 30 days by email hash
• Exclude calls shorter than 60 seconds unless your close rate data proves otherwise
• Exclude internal traffic and partner test leads
• Separate content downloads from sales-intent forms entirely

Now consider a concrete example. A campaign reports 120 leads in Google Ads. After cleanup:

• 18 are duplicates
• 22 use blocked domains or nonsense inputs
• 27 are content-download forms
• 53 are valid sales leads

Your usable conversion rate is not based on 120. It is based on 53. If spend was €9,600, your reported CPA is €80, but your clean CPA is €181.13.

That single cleanup step often changes bidding decisions, landing page priorities, and keyword strategy all at once. It also makes return calculations far more honest, because your numerator stops pretending every lightweight lead has equal value.

When should you import offline conversions?

Whenever the meaningful business event happens after the click. This is especially relevant in B2B, high-ticket services, and any long sales cycle. WordStream’s 2026 conversion tracking guide notes that uploaded offline conversions can be imported through a CRM platform in Google Ads. If the real value signal is SQL, opportunity created, or customer won, then relying only on website events is too shallow.

Here is a worked example with a simple weighting system:

• Demo submitted: value 20
• Sales qualified: value 60
• Opportunity created: value 120
• Customer closed: value 300

Suppose keyword group A generates:

• 30 demos = 600 points
• 9 SQLs = 540 points
• 3 opportunities = 360 points
• 1 customer = 300 points
• Total value score: 1,800

Keyword group B generates:

• 18 demos = 360 points
• 12 SQLs = 720 points
• 5 opportunities = 600 points
• 2 customers = 600 points
• Total value score: 2,280

If you optimise only to website demos, group A looks stronger. If you import downstream events, group B is the better business driver.

Once the event structure is clean, the next weak point is usually the transport layer. That is where client-side setups start to collapse.

Use server-side tracking first

Client-side-only tracking used to be “good enough” for many advertisers. It is not a safe default anymore. HubSpot Community’s 2025 reply in the 2024 tracking thread says that client-side Google Tag Manager tracking is becoming increasingly unreliable because of browser privacy changes, ITP, and ad blockers, and recommends server-side tagging combined with a robust API-based approach as the better modern solution. That matches what many serious teams now see in practice: browser-executed scripts fail too often to act as the sole source of truth.

This is the technical core of google ads conversion tracking accuracy. If the event never reaches the platform, no attribution model can save it.

Why is client-side GTM breaking?

The browser is now an adversarial environment for ad measurement. Safari’s privacy protections, ad blockers, consent interruptions, script load failures, and cross-domain handoff issues all increase drop-off. A perfectly configured client-side tag can still miss the event because the browser never lets it complete.

That is why Forrester’s 2024 cookie deprecation analysis matters beyond remarketing. Forrester argues that data deprecation will create a measurement gap by weakening the signals Google uses to tie ad exposures to conversion events. If identity and browser-based signal continuity degrade, client-side fragility becomes a business risk, not just a technical annoyance.

A practical symptom list includes:

• Click volumes look normal, but mobile conversions fall sharply on privacy-heavy browsers
• CRM leads exceed ad-platform conversions by a wide margin
• Form backend counts are higher than analytics event counts
• Consent banner interactions cause event underreporting in certain regions

The contrarian point is worth stating plainly: more client-side tags do not solve client-side fragility. They usually multiply it.

What does server-side tracking actually fix?

Server-side tracking does not create perfect truth, but it removes several common breakpoints. It captures the event from your server or cloud container after the business action completes, rather than hoping the user’s browser sends every script successfully. That improves reliability, supports better deduplication, and creates a cleaner bridge to platforms and CRMs.

Our preferred Signal Survival Stack has three layers:

1. Clean event design: one primary conversion, controlled secondary actions, explicit dedupe rules
2. Server-side capture: send conversion events from a trusted backend path, not only the browser
3. CRM reconciliation: match reported conversions to qualified leads, opportunities, and customers

The framework matters because each layer protects against a different failure mode. Event design prevents junk learning. Server-side capture reduces browser loss. CRM reconciliation catches attribution drift.

Here is a scenario with numbers. Before a server-side setup:

• Form backend submissions: 84
• Google Ads conversions: 51
• CRM leads: 79

After server-side event forwarding with dedupe:

• Form backend submissions: 86
• Google Ads conversions: 76
• CRM leads: 81

No system reaches perfect parity, but the gap drops from 33 missing conversions to 10. That is a material improvement in bidding input.

How do you implement a server-side setup without chaos?

Keep the rollout boring. The biggest mistake is rebuilding the entire stack at once and losing comparability. A controlled migration looks like this:

• Keep the existing client-side event live temporarily
• Add a server-side event with a shared event ID for deduplication
• Compare browser, server, and backend counts for 2-4 weeks
• Move bidding to the cleaner primary event only after the variance stabilises
• Document every field passed: click IDs, timestamps, consent status, source, lead type

If you also run landing page tests, pair this with a disciplined experimentation setup. Teams often mistake measurement changes for conversion-rate gains. That is one reason we recommend reading our guide on testing tools that can support cleaner validation when you redesign the stack.

Even with server-side improvements, the macro environment is getting noisier. That is the next structural challenge.

Prepare for the cookie gap

Forrester’s 2024 analysis of cookie deprecation makes two points that advertisers should take seriously. First, third-party cookie loss will hamstring remarketing lists for search ads by reducing data on behaviour across non-Google properties. Second, data deprecation will weaken the signals Google uses to connect ad exposures to conversion events, which creates a broader measurement gap. This is not a temporary implementation issue. It is a structural shift.

That matters because many Google Ads accounts still rely on borrowed identity. They assume the platform can infer enough cross-site behaviour to maintain targeting and attribution quality. That assumption is deteriorating.

What happens when cross-site signals disappear?

You lose both reach and confidence. Remarketing pools become thinner, attribution paths get noisier, and smart bidding has less context. Forrester recommends prioritising zero-party and first-party data and investing in microexperiences that generate loyalty, recommendations, or exclusivity. In practical Google Ads terms, that means creating more moments where users willingly identify themselves or move into known audience paths.

A simple example: suppose your remarketing list used to include 20,000 site visitors per month, and cookie loss reduces matchable users by 35%. Your reachable pool drops to 13,000. If your historical retargeting conversion rate was 4.5%, the expected monthly conversion volume from that audience can shrink materially even if nothing else changes.

Now compare two strategies:

• Old model: broad site retargeting based on browser identifiers
• New model: persona-specific landing experiences with form capture and enriched first-party data

The second does not fully replace lost scale, but it creates a more durable signal.

Why remarketing lists get worse, not better

A common assumption is that Google will “figure it out” with automation. Sometimes it compensates. Often it obscures the problem. The more your account depends on cross-site observation, the more your measurement quality becomes platform-dependent and less auditable.

That is why we push teams away from pure audience dependency and toward path dependency. If you can guide a visitor through a clear intent sequence on your own property, you need less inferred identity from third parties.

Think of two landing page experiences for the same keyword set:

• Version A: one generic page for all visitors, one short form, no branching
• Version B: persona selector, dynamic message block, tailored CTA, segmented form routing

If Version B converts fewer total leads but produces more reliable source data and more qualified outcomes, it is often the better measurement asset. Accuracy beats superficial volume when the goal is better bidding and better revenue decisions.

When should you invest in first-party data capture?

Earlier than most teams think. Not when attribution has already degraded beyond trust. The trigger is simpler: if paid search meaningfully drives pipeline, you need a first-party measurement plan now.

That plan usually includes:

• Persistent click ID capture where consent allows
• Hidden-field source stamping on forms
• Email hashing or equivalent privacy-safe identifiers for enhanced matching
• Offline conversion imports tied to downstream CRM status
• Persona-path analytics on landing pages

This links directly to Forrester’s recommendation to invest in experiences that encourage users to self-identify. It also connects to the broader reality that HubSpot’s 2026 marketing statistics page highlights: lead-to-customer conversion ranks as the second most important KPI for marketers, and CRO is the second-most-used optimisation technique at 50%. Teams already know they need better conversion efficiency. The next step is designing measurement systems that make those optimisation efforts trustworthy.

And that takes us to the part many tracking guides ignore: the page itself.

Build landing pages that measure cleanly

Forrester’s 2024 guidance on marketing to anonymous website visitors says most website visitors are anonymous and do not take action to identify themselves. Their recommendation is not vague. They advise using predefined personas, dynamic content in a WCMS, and conditional logic to map visitors into persona-specific conversion paths, then measuring those paths against non-path experiences. That is an important shift. Better tracking is not only about instrumentation. It is also about making intent legible on the page.

When every ad click lands on one generic page, attribution gets muddy. You do not know whether poor performance comes from traffic quality, message mismatch, or weak UX. Persona-based landing paths reduce that ambiguity.

How do you track anonymous visitors?

You do not wait for perfect identification. You infer intent from controlled choices. A good landing page for anonymous traffic asks the visitor to reveal context through actions that also improve conversion routing.

Examples include:

• “I’m evaluating for my team” vs “I’m an agency”
• “I need help with landing pages” vs “I need help with Google Ads reporting”
• “Budget under €5k/month” vs “Budget over €25k/month”

These are not just qualification fields. They are measurement enrichers. Once a visitor selects a path, you can compare path conversion rates, qualified lead rates, and closed-won rates by path.

Forrester’s recommendation to measure predefined paths against non-path experiences gives you a practical testing model. It also aligns with our broader view that landing page design and tracking design should be one conversation, not two separate workstreams.

Should every ad send traffic to the same page?

Usually not. A single page can work for tightly grouped intent. It breaks down when audiences, offers, and buyer stages diverge. That is especially true in B2B search, where the same account may target competitor terms, problem-aware terms, and solution-aware terms with very different expectations.

Here is a simple comparison:

A concrete scenario makes the trade-off clearer. Suppose a campaign group targeting competitor keywords sends visitors to the same page as branded traffic. The page converts at 7%, but only 20% of those leads become SQLs. A dedicated competitor-comparison page converts at 5.5%, yet 45% become SQLs. Lower top-line conversion rate, higher measurement quality, better business result. That is why generic page volume can be deceptive. If competitor traffic is part of your mix, our guide to tracking rival ad strategy is useful context for aligning intent and page paths.

The Path Clarity Framework

The Path Clarity Framework is our landing-page companion to tracking accuracy. It has three steps: match the ad intent, force an informative choice, and tie each path to a distinct primary conversion. The goal is not more complexity. It is cleaner diagnosis.

Consider a company spending €12,000/month across three keyword clusters:

• Brand terms: 400 clicks, generic page, 12 conversions, 8 SQLs
• Problem-aware terms: 900 clicks, generic page, 27 conversions, 9 SQLs
• Competitor terms: 500 clicks, generic page, 18 conversions, 3 SQLs

Now redesign the landing experience:

• Brand terms stay on a straightforward demo page
• Problem-aware traffic gets an industry-selector path
• Competitor traffic gets a comparison page with a qualification step

After 30 days:

• Brand terms: 13 conversions, 9 SQLs
• Problem-aware terms: 24 conversions, 12 SQLs
• Competitor terms: 12 conversions, 5 SQLs

Total conversions fall from 57 to 49, but total SQLs rise from 20 to 26. That is the kind of trade-off most dashboards hide if they only count front-end form fills.

The edge case is obvious: not every account needs elaborate pathing. If spend is low and keyword intent is tightly controlled, a simpler page may outperform through sheer focus. But once multiple intents and buyer types share the same destination, page architecture becomes part of measurement hygiene.

Once pages collect cleaner intent and your server-side setup preserves the events, you still need one more discipline: reconciliation.

Use a measurement stack you trust

A trustworthy Google Ads setup does not ask one platform to be the source of truth for everything. HubSpot Community’s tracking discussion points toward the right operating model: use ad attribution reporting, but also mark contacts as customers as they actually become customers. In other words, combine platform data, server-side events, CRM status, and landing page path performance in one reconciliation loop.

This is the final piece of the thesis. Conversion tracking accuracy is a system problem. If you only read one system, you will inherit its blind spots.

What should your source of truth be?

For bidding, your source of truth should be the cleanest timely signal that predicts revenue. For finance and performance review, your source of truth should be the CRM or downstream customer system. Those are not the same thing, and they should not be.

We usually recommend this split:

• Google Ads primary conversion: a high-intent, deduped event such as demo request or qualified call
• Analytics and server logs: validation layer for event integrity
• CRM: business truth for SQLs, opportunities, customers, and revenue
• Landing page reporting: path and intent diagnostics

This setup matters because speed and certainty are different jobs. Google Ads needs a timely optimisation signal. Leadership needs a financially meaningful one.

How often should you audit tracking?

More often than quarterly. In active accounts, we recommend a weekly light audit and a monthly full reconciliation.

A light audit checks:

• Are primary conversion counts moving in line with backend submissions?
• Did any browser or device segment drop sharply?
• Are duplicate rates or spam rates rising?
• Did consent or form changes affect capture?

A full monthly reconciliation checks:

• Google Ads primary conversions vs server-side event counts
• Server-side event counts vs CRM lead creation
• CRM leads vs qualified leads and customers
• Conversion rate by landing page path and keyword intent cluster

Here is a practical threshold model:

The contrarian point here is useful: you do not need perfect attribution to make better decisions. You need consistent, auditable, decision-grade measurement. Teams waste months chasing impossible precision while ignoring obvious event pollution.

The reconciliation loop in practice

Suppose an account shows this monthly picture:

• Google Ads conversions: 140
• Server-side primary events: 154
• CRM leads: 148
• SQLs: 52
• Customers: 11
• Spend: €18,000

At first glance, the account-level CPA is €128.57 based on Google Ads conversions. But the more decision-useful numbers are:

• Cost per CRM lead: €121.62
• Cost per SQL: €346.15
• Cost per customer: €1,636.36

Now add path reporting and you may find that one landing page cluster produced 60 conversions but only 12 SQLs, while another produced 38 conversions and 20 SQLs. That changes budget allocation immediately.

This is also where adjacent channel context helps. If your cost structure looks strange, benchmarking against broader paid media economics can help frame whether the problem is tracking, traffic, or market pricing. Our article on cost-per-lead ranges by industry can be a useful reference when sanity-checking the output of your measurement system.

A reliable stack does not eliminate uncertainty. It contains it. And once the system is stable, optimisation work starts compounding instead of resetting every time a tag breaks.

Turn accuracy into better campaigns

If your Google Ads reports are unstable, dynares.ai helps you fix the system, not just the symptom. We combine persona-based landing page generation, conversion-focused page testing, and measurement-aware optimisation so the signal Google Ads sees is cleaner from the first click to the CRM handoff. That matters when you are dealing with the exact problems covered here: weak page-path clarity, noisy front-end events, and campaigns that look unprofitable because tracking undercounts or misclassifies conversions.

Our platform is built for teams that want tighter control over landing page paths, faster iteration on message-match pages, and better alignment between ad intent and conversion design. If you are already investing in paid search, this gives you a practical way to stop sending mixed-intent traffic to generic pages and stop making budget decisions on shaky data. You can explore how dynares.ai supports cleaner conversion journeys, more testable page experiences, and a measurement stack that is easier to trust when privacy loss keeps increasing. The teams that win over the next year will not be the ones with the most dashboards. They will be the ones that build conversion signals durable enough to survive reality.