Marketers love to say we’re “data-driven.” It makes us sound responsible, like we wear lab coats and don’t make life-altering decisions based on vibes.
In reality, we’re more like fortune tellers with dashboards.
But here’s the thing: the problem isn’t that we’re not data-driven. It’s that we’re treating data like decoration instead of direction. Let’s fix that.
1. The Spreadsheet Mirage (Or: Why Your Data Is Lying to You)
Every marketer has one sacred spreadsheet. It’s 47 tabs long, color-coded like a toddler’s art project, and updated religiously until week two of the quarter, when everyone quietly switches to guessing.
We say things like “Let’s circle back when the numbers mature.” The numbers are not wine. They’re barely juice.
The Survival Strategy: Track Less, Track Better
Your problem isn’t too little data. It’s too much of the wrong data and not enough of the right data.
Your Data Diet:
Stop tracking:
- Metrics you never actually use to make decisions
- Vanity metrics that feel good but don’t change behavior (looking at you, “total impressions”)
- Things you can’t influence (market trends, competitor actions—track them, but don’t put them in YOUR performance dashboard)
- Data that requires 3 calculations and 2 pivot tables to understand
Start tracking:
- Metrics that directly connect to revenue (leads generated, conversion rate, customer acquisition cost)
- Metrics you can actually improve week-over-week (email open rates, landing page conversion rates, form completion rates)
- Leading indicators, not just lagging ones (website traffic trends predict future leads better than last month’s closed deals)
Your New Spreadsheet Rules:
✓ One source of truth – Not 47 tabs. One dashboard that everyone looks at. ✓ Update it or kill it – If you haven’t updated a metric in 30 days, delete it. ✓ The “so what?” test – For every metric, answer: “If this number changes, what do I do differently?” If you can’t answer, you don’t need that metric.
Real talk: I cut my reporting from 30 metrics to 5. Turns out, when you’re only tracking things that matter, you actually pay attention to them.
2. The A/B Test That No One Believes (Or: How to Run Tests That Actually Mean Something)
We test subject lines like “Open This Email” vs. “Open This Email Now.” One wins by 0.3%, and suddenly someone in upper management says, “We should build the entire campaign strategy around this insight.”
Sure, Jan. Let’s base a six-figure decision on 12 people who didn’t hit spam.
The Survival Strategy: Test Things That Matter, At Scale That Matters
The problem isn’t A/B testing. It’s testing tiny things with tiny sample sizes and treating the results like gospel.
Your Better Testing Framework:
What’s worth testing:
- Different value propositions (not just word tweaks)
- Different audience segments (who converts better?)
- Different page structures (long-form vs. short, video vs. text)
- Different offers (free trial vs. demo vs. download)
What’s not worth testing (yet):
- Button colors when your conversion rate is under 1%
- Subject line punctuation when your open rate is 10%
- Tiny copy changes when you haven’t validated your core message
Your Testing Hierarchy:
- Test big strategic things first (message, offer, audience)
- Once those work, test tactical things (headlines, layouts, CTAs)
- Only then test the polish (colors, button text, image placement)
The statistical reality check:
For a test to be meaningful:
- You need at least 100 conversions per variation (not visits, CONVERSIONS)
- You need at least 95% statistical confidence
- You need to run it for at least one full business cycle (week, month, whatever makes sense)
If your test doesn’t hit these bars, you’re not testing. You’re guessing with extra steps.
How we actually do this:
In our own work, we use Leadpages specifically for testing because we can spin up multiple landing page variations in under an hour, and their built-in A/B testing actually tracks statistical significance. We’re not sitting there with a calculator trying to figure out if a 2% lift is real or random noise.
We test big stuff: different headlines that reframe our value prop, different page structures, different offers. Then we let it run until we have real data (usually 2-3 weeks, minimum). If we tested button colors before we validated our core message, we’d be wasting everyone’s time.
The rule: Don’t test unless you’re ready to act on the results. And don’t act on results that aren’t statistically sound.
3. The Persona Problem (Or: Your Ideal Customer Is Real, You Just Haven’t Talked to Them)
“Meet Olivia, the 34-year-old marketing manager who loves kombucha and SaaS webinars.” We built a whole campaign for her. She’s not real. She’s an intern’s fever dream in Canva.
Meanwhile, actual customers are out there just trying to find the ‘unsubscribe’ link.
The Survival Strategy: Build Personas From Real Humans, Not Assumptions
The problem with personas isn’t the concept. It’s that we make them up instead of discovering them.
Your Real Persona Research Process:
Step 1: Talk to 10 actual customers Ask them:
- What problem were you trying to solve when you found us?
- What almost stopped you from buying?
- What convinced you to buy?
- How would you describe us to a colleague?
Step 2: Look at your actual data
- Which customer segments have the highest lifetime value?
- Which acquisition channels bring the best customers?
- What content do your best customers consume?
Step 3: Find the patterns You’re not looking for demographic details (“34-year-old marketing manager”). You’re looking for psychographic truths:
- What are they worried about?
- What do they value?
- What language do they use?
- What objections do they have?
Step 4: Write personas that are actually useful
Bad persona: “Marketing Manager Olivia, 34, lives in Austin, drinks kombucha, 2 kids”
Good persona: “Stretched-Thin Sarah is responsible for demand gen but has no budget for agencies. She needs tools that work immediately without a learning curve. Her biggest fear is investing time in something that doesn’t show ROI within a quarter. She makes decisions based on peer recommendations and free trials.”
See the difference? One is creative writing. The other is a roadmap for messaging, positioning, and product.
Reality check: If your persona includes their favorite beverage but not their biggest business problem, start over.
4. Attribution: The Modern Ghost Story (Or: How to Track What Actually Matters)
Nobody knows where the leads actually came from. Was it the webinar? The ad? That one tweet someone’s intern posted at 2 AM? We don’t know, but we’ll still make a 27-slide deck pretending we do.
Marketing attribution is basically Schrödinger’s funnel: it’s working and not working at the same time.
The Survival Strategy: Embrace Imperfect Attribution (But Make It Useful)
Perfect attribution is a myth. But “good enough to make smart decisions” attribution is totally achievable.
Your Practical Attribution System:
What you CAN track:
- First touch (what brought them in?)
- Last touch (what closed them?)
- Direct traffic trends (are people typing in your URL? That’s brand awareness working)
- Campaign-specific landing pages (did THIS specific campaign drive conversions?)
What you CAN’T perfectly track:
- That podcast they listened to three months ago
- The LinkedIn post they saw but didn’t click
- The word-of-mouth recommendation from their colleague
- The “dark social” links from Slack or text messages
Your action plan:
✓ Use UTM parameters religiously – Every link, every campaign, every time. No exceptions.
✓ Use campaign-specific landing pages – Don’t send all traffic to your homepage. Create unique pages for each campaign so you can actually see what’s working. (This is why we use dedicated landing pages for every major campaign—we can track exactly which message, which offer, which channel drove conversions.)
✓ Survey your customers – Add one question to your onboarding: “How did you first hear about us?” You’ll get insights attribution software misses.
✓ Track time-to-conversion – Leads that convert in 24 hours = probably that last ad they clicked. Leads that convert in 30 days = probably multiple touchpoints including brand awareness efforts.
✓ Accept the gray area – Some credit goes to brand. Some goes to performance. Some goes to luck. Model it the best you can, then make decisions based on trends, not individual data points.
The reality: Your webinar probably didn’t “cause” all those conversions. But if leads who attend webinars convert 3x more than leads who don’t, keep doing webinars. You don’t need perfect attribution. You need directionally correct decisions.
5. KPIs and Other Fantasies (Or: The Metric That Won’t Die)
We all have that one KPI that never moves. It’s been red for three quarters, but no one dares delete it because it’s “important.”
Like a haunted house, we just live with it. “Ignore the screaming metric, it’s part of the ambiance.”
The Survival Strategy: Kill Your Zombie Metrics
If a metric has been red for three quarters and nobody’s done anything about it, it’s not a KPI. It’s a wish.
Your Metric Audit:
Go through every KPI you track and ask:
✓ Is this a real priority? – If it’s not in your top 3 goals, it’s not a KPI.
✓ Can we actually influence it? – If you can’t change it through your actions, it’s context, not a KPI.
✓ Are we taking action based on it? – If it’s red and nothing changes, delete it.
✓ Is this a leading or lagging indicator? – Leading indicators (website traffic, email signups) predict future success. Lagging indicators (revenue, closed deals) tell you what already happened. You need both, but focus on leading indicators for week-to-week decisions.
Your New KPI Framework:
The Rule of 3-5-1:
- 3-5 total KPIs (not 20)
- 1 North Star Metric that matters most (probably revenue, or whatever leads to revenue)
- Everything else is supporting data, not a KPI
Example:
North Star: Monthly Recurring Revenue (MRR)
Supporting KPIs:
- Marketing Qualified Leads (leading indicator of future MRR)
- Lead-to-Customer Conversion Rate (are we attracting the RIGHT leads?)
- Customer Acquisition Cost (are we acquiring customers efficiently?)
- Customer Lifetime Value (are we acquiring VALUABLE customers?)
That’s it. Four metrics that tell you: Are we growing? Are we attracting the right people? Are we doing it efficiently? Are they valuable?
The hard truth: If your dashboard has 15 red metrics, you don’t have 15 problems. You have 1 problem: you’re not prioritizing.
Pick the 3 things that matter most. Make them green. Then pick the next 3.
Final Thought: Be Data-Driven, Not Data-Decorated
Here’s the difference:
Data-decorated marketing:
- 47-tab spreadsheets nobody reads
- A/B tests with no statistical significance
- Made-up personas
- Attribution models that require a PhD to understand
- KPIs that never change anything
Data-driven marketing:
- 5 metrics that everyone understands and acts on
- Tests that matter, with sample sizes that matter
- Personas built from real customer conversations
- Attribution that’s “good enough” to make smart bets
- KPIs that actually drive decisions
You don’t need more data. You need better questions.
So yes, we’re all still a little bit like fortune tellers with dashboards. But at least now we’re using the right crystal ball.
May your tests reach statistical significance, your attribution be directionally correct, and your KPIs actually be key.
Now go track something that matters.
