The pattern nobody wants to admit: we’re drowning in output, starving for signal
Scan the headlines you just read and a pattern jumps out:
- “Most Marketing Metrics Are Misleading.”
- “11 Best Social Media Analytics + Reporting Tools.”
- “How Our Website Conversion Strategy Increased Business Inquiries by 37%.”
- “Google launches developer hub for ads and measurement tools.”
- “Marketing forecast fundamentals every growth team needs.”
- “AI’s trust problem: The cost of outsourcing your message.”
The industry is obsessed with more: more tools, more dashboards, more metrics, more AI. But the operators who are actually winning are doing something much less glamorous:
They are aggressively shrinking what they measure.
Not because they’re anti-data, but because they finally realized the uncomfortable truth:
most “performance” reporting is just a very expensive way to avoid answering one question:
Is this channel actually compounding profitable growth, or just creating busywork?
Why most “performance” teams are really “activity” teams
Look at your own operation for a moment:
- Paid team reports on ROAS, CTR, CPC, view rate, VTR, thumb-stop rate.
- SEO team reports on rankings, impressions, clicks, cannibalization, title tags.
- Social team reports on followers, engagement rate, saves, shares, reach.
- CRM team reports on open rate, click rate, unsubscribe rate, spam rate.
Every team can “prove” they’re doing well inside their own sandbox. Yet:
- Incremental revenue is flat.
- Blended CAC is creeping up.
- Payback periods are getting longer.
- Sales says “leads are worse than last year.”
That’s the KPI trap: each function is optimized to look good on its own scoreboard, while the business is quietly underperforming.
The new reality: AI makes bad metrics look better than ever
AI is pouring gasoline on this problem:
- AI writing tools can crank out SEO content that ranks on long-tail terms but never moves revenue.
- AI video editing can double your content output while your watch time and sales don’t budge.
- AI ad systems can optimize for “conversion” events that are cheap but low value or non-incremental.
- AI analytics tools can generate gorgeous dashboards that tell you… nothing new.
When your stack can auto-generate content, auto-bid, auto-optimize, and auto-report, the risk isn’t that you’ll have too little data. It’s that you’ll have too many good-looking lies.
The job of a modern CMO or performance lead is no longer “get more data.” It’s:
decide which 90% of data you’re willing to ignore.
The only three questions your metrics should answer
At the company level, almost everything you measure should ladder into three questions:
- Are we buying profitable growth?
That’s revenue, contribution margin, and fully loaded CAC by cohort, not just by channel. - Is this channel incrementally valuable?
That’s lift, not last-click. Think experiments, geo splits, and incrementality tests. - Are we compounding or just renting attention?
That’s retention, LTV, and owned audience growth, not just this month’s ROAS.
If a metric doesn’t help you answer one of those three, it’s either:
- a diagnostic metric (useful for operators, not for steering the business), or
- a vanity metric (useful for slideware, not for decisions).
From “metrics buffet” to “measurement spine”
High-output teams that actually perform share one thing: a ruthless, boring measurement spine that everything else plugs into.
Here’s what that looks like in practice.
1. Start with a spine of five company-level metrics
Make these non-negotiable, published weekly, and shared across marketing, product, and finance:
- New revenue by cohort (or new customers by cohort) with contribution margin.
- Blended CAC (total sales and marketing spend / new customers).
- Payback period on CAC by major channel.
- 12-18 month LTV by acquisition cohort and channel.
- Incremental lift from major paid channels (updated quarterly via experiments).
Everything else is secondary. Not unimportant, but subordinate.
2. Force every channel to “speak” in the same currency
Most teams still report in their native tongue:
- SEO: “We grew organic traffic 40%.”
- Paid social: “We’re at 2.5x platform ROAS.”
- Social: “Engagement is up 30%.”
- CRM: “Open rates are up 5 points.”
None of that is comparable, which means you can’t make real tradeoffs.
The fix: standardize on contribution margin and payback period as the shared language.
- SEO doesn’t just report traffic; it reports incremental revenue and margin from organic vs. a synthetic “no-SEO” baseline.
- Paid social doesn’t just report platform ROAS; it reports experiment-based lift and modeled payback.
- Social and content don’t just report engagement; they report assisted revenue and owned audience growth tied to LTV.
- CRM doesn’t just report open rate; it reports incremental revenue vs. holdout and churn reduction.
3. Treat platform metrics as suspects, not facts
Google claims its AI-powered ads can lift sales by 80%. OpenAI is building an ads business. Every platform is now grading its own homework with AI.
If you treat those numbers as truth, you’re not a performance marketer; you’re a reseller.
Instead:
- Run structured incrementality tests:
geo holdouts, audience holdouts, time-based on/off tests. - Use MMM or lightweight Bayesian models as a sanity check on platform-reported conversions.
- Compare “always-on” vs. “pulsed” spend to see if the curve is actually concave or if you’re just paying for organic demand.
- Align with finance on what counts as a “real” conversion (refunds, cancellations, and fraud removed).
Cleaning up the metric mess: a practical playbook
Here’s how to move from “metrics buffet” to a spine in 90 days without blowing up your operation.
Step 1: Run a ruthless metric audit
In each weekly or monthly report, ask:
- What decision did this metric change in the last 60 days?
- If we stopped reporting it, what would actually break?
- Is this a steering metric (changes budget/strategy) or a diagnostic metric (helps operators debug)?
Kill or demote anything that:
- isn’t tied to a decision, or
- can’t be causally connected to revenue, margin, or LTV.
Step 2: Collapse channel scorecards into business scorecards
Replace channel-specific decks with a single cross-channel view:
- Rows: channels (search, paid social, organic, email, partnerships, etc.).
- Columns: spend, new customers, incremental revenue, contribution margin, payback, LTV/CAC.
Channel teams still keep their diagnostic dashboards, but the exec view is unified and brutally simple.
Step 3: Rebuild goals around fewer, harder metrics
If your teams are bonused on soft or channel-native metrics, they will optimize for them. That’s how you get:
- SEO teams celebrating traffic spikes from keywords that never convert.
- Media buyers scaling campaigns that “look great” in-platform but don’t pass an incrementality test.
- Social teams chasing followers that never join your email list or buy.
Shift incentives to:
- Channel-level CAC vs. target (with quality controls from sales or product).
- Payback within X months by channel or initiative.
- Incremental LTV for cohorts acquired through specific plays.
Yes, it’s harder to measure. That’s the point. Easy metrics are usually lying.
Step 4: Put AI on a leash
AI should compress time, not judgment. Use it to:
- Generate hypotheses faster (creative variations, audience ideas, keyword clusters).
- Automate low-level reporting and anomaly detection.
- Summarize qualitative data (reviews, comments, sales calls) into patterns.
Don’t let it:
- Define success (that’s your job, via the measurement spine).
- Pick optimization events without human review.
- Produce content at scale without guardrails on brand, accuracy, and performance.
AI is a force multiplier. If your metrics are bad, it will just help you go off the cliff faster.
What “high-signal” actually feels like for operators
When you get this right, the day-to-day feels very different:
- Your weekly growth meeting is 2-3 pages, not 40 slides.
- Media buyers argue about incremental CAC, not CTR.
- SEO talks about new revenue from high-intent queries, not “total keywords ranked.”
- Social reports on email signups and LTV from content, not just saves and shares.
- Finance doesn’t roll their eyes at marketing’s numbers.
You’ll still have disagreements. You’ll still be wrong sometimes. But you’ll be wrong about the right things: pricing, positioning, creative, channel mix – not whether a 0.3% change in click-through rate is “good.”
How to start this quarter, not “after we replatform”
You don’t need a new CDP, a clean-room, or a 12-month data project to fix this. You can start in the next sprint:
- Pick your spine: agree on 4-6 company-level metrics that matter this year.
- Rewrite your decks: every recurring marketing and growth deck starts with that spine.
- Pick one channel and run a proper incrementality test in the next 30 days.
- Kill three metrics from your next report that nobody will miss.
- Audit AI usage: for each AI tool, write down what decision it’s influencing and how you’ll know if it’s wrong.
High-signal marketing isn’t about having the most data. It’s about having the fewest excuses.