Every marketer claims to be "data-driven". Most are data-adjacent at best. They've got GA4 installed, they glance at it on a Monday, and they call that strategy. It isn't. Proper analytics changes how decisions get made, not just what gets reported.
The difference between reporting and analysis
Reporting is "here's what happened last month". Analysis is "here's why it happened, and here's what to do next". Most agencies sell you the first and charge you for the second. That's a shame, because without the "why" the numbers are just a receipt.
If your monthly report doesn't end with a clear decision, a test to run, a budget to shift, a page to fix, it's not analytics. It's scrapbooking.
The metrics that actually move the needle
- Customer acquisition cost (CAC): what it costs to get one paying customer, end to end, not just the ad spend.
- Customer lifetime value (LTV): what that customer's worth over their relationship with you.
- LTV:CAC ratio: the single most important number in your business. Under 3:1, you're working too hard. Over 5:1, you're underinvesting in growth.
- Payback period: how long until a new customer's generated back what they cost to acquire.
Everything else, impressions, clicks, engagement, bounce rate, is a leading indicator of those four. Leading indicators matter, but they serve the headline numbers, not the other way round.
Attribution: a polite fiction
Everyone wants to know which channel drove the sale. The honest answer is usually "all of them, a bit". A customer might see a TikTok ad, google you a week later, read a blog, subscribe, get three emails, then finally buy after a retargeting ad. Which channel gets the credit?
Last-click attribution is still the default in many tools, and it's lying to you. It gives all the credit to the final touchpoint and none to the channels that built awareness. Position-based or data-driven attribution gets closer to reality, but even then, it's directional, not gospel.
The fix isn't a better attribution model. It's incrementality testing: turn a channel off, see what happens to revenue. Brutal but honest.
Cohort analysis: the quiet superpower
Aggregate numbers hide stories. Cohort analysis, grouping customers by when they joined and tracking them over time, reveals whether your product's genuinely improving or whether you're just running faster to stand still.
If customers acquired in March have a higher 90-day retention than those from January, something's working. If it's the opposite, something's broken, and no amount of top-of-funnel spend will fix it.
Segmentation before averages
Averages are the enemy of insight. "Our conversion rate is 2.1%" means nothing. 2.1% across what? Paid vs organic, mobile vs desktop, new vs returning, Newcastle vs London, all perform differently. The moment you segment, patterns appear.
We had a client whose mobile conversion rate was a third of desktop. Average looked fine. Split out, it was a five-figure monthly leak hiding in plain sight. Fixed the mobile checkout, recovered the gap inside six weeks.
The tool stack, kept sensible
- GA4: free, powerful, annoying. Set it up properly or it'll mislead you.
- Looker Studio: for actually looking at the data instead of clicking round GA4's menus.
- Hotjar or Microsoft Clarity: session recordings show you what the numbers can't.
- Server-side tagging: increasingly necessary as browsers tighten cookies.
You don't need a £5,000-a-month analytics suite. You need someone who'll use GA4 like they mean it.
Common analytics mistakes
Everyone makes these. We've made them. We try not to anymore.
- Tracking everything, analysing nothing: more data doesn't mean more insight.
- Chasing statistical significance on tiny samples: a 50-visitor A/B test is astrology.
- Ignoring qualitative data: customer interviews beat charts for understanding "why".
- Reporting vanity metrics to clients: everyone feels good, nobody grows.
Turning data into action
Every analytics session should end with one of three outcomes: a decision made, a test set up, or a question refined. If it ends with "interesting, let's look again next month", you've wasted an hour.
For specifics on how this plays out in search, SEO competitor analysis walks through one practical application. For ads, Google Ads performance metrics covers the same ground.
Drowning in dashboards, short on answers?
We'll untangle your analytics, focus on the numbers that matter, and give you a report that actually tells you what to do next.
Book an analytics reviewFinal thought
Data doesn't make decisions. People do. Analytics just means those people are deciding from evidence instead of vibes. Use it well and you'll outgrow competitors twice your size. Use it badly and you'll have very pretty charts of a business going nowhere.
