Data driven marketing means deciding where the budget goes by looking at what the money did last time. It is a habit rather than a tool. This article covers which numbers deserve to move money, and when they cannot settle a question at all. It also covers the ways a number quietly misleads you.
Most budget meetings do not work like that. Somebody defends the channel they like, somebody else defends the one they run, and the loudest voice wins. A shared set of numbers ends that argument before it starts.
What data driven marketing actually means
It does not mean more dashboards. Plenty of businesses already own more charts than anybody reads.
Data driven marketing means one thing. A number changes a decision. If a report arrives, gets a nod, and changes nothing, then it is decoration. Consequently the test is not how much you measure. The test is how often a measurement moves money.
That habit sits underneath marketing that has to pay for itself. A channel keeps its place only while somebody checks what it returned. So the useful question at the end of any report is short. What are we doing differently next month, and which number told us to?
Start from the decision, not the dashboard
Most reporting projects begin at the wrong end. Somebody connects every tool, builds a page of charts, and then wonders which of them matters.
Work backwards instead. Write down the decisions you actually make. Then find the smallest number that would settle each one.
- Should we spend more here or less? Cost per genuine enquiry answers that.
- Is this channel bringing buyers or browsers? Look at enquiry quality instead.
- Do we need another channel? Compare current volume against the work you can take.
- Which page loses the click we paid for? Enquiry rate by page shows it.
Four decisions need four numbers. Everything else can live in an appendix, where it does no harm and asks nothing of anybody.
Choosing that short list is a job in itself. Our guide to marketing numbers that change decisions goes through the usual candidates one at a time.
The numbers that deserve to move money
A number earns the front page when it survives one question. If this moved, would we act?
Cost per enquiry passes easily. Impressions usually fail, because nobody moves a budget over them. Meanwhile the numbers in between are the interesting ones, and each hides something.
| The number | What it hides |
|---|---|
| Clicks | Cheap clicks are easy to buy, so the total climbs while interest falls. |
| Cost per lead | It drops when you attract people who will never buy. |
| Return on ad spend | Revenue is not profit, so a healthy figure can still lose money. |
| Website traffic | It can grow while enquiries stay flat. |
None of those is useless. Each one becomes dangerous only when somebody promotes it to a headline and stops asking what sits underneath.
Return on ad spend is the easiest of the four to misread, since revenue hides the cost of delivering the work. Before you defend a budget with it, read why the figure flatters you.
Where data driven marketing goes wrong
The failures are rarely about arithmetic. Usually somebody reads a number as more certain than it is.
Small numbers say very little
A month with four enquiries cannot tell you which channel wins. One extra enquiry would flip the ranking. Yet plenty of budgets move on exactly that.
When volumes are low, widen the window. Three months beat one, and a year beats three. Until you have enough, treat the figures as a hint rather than an answer.
Attribution is a model, not a fact
Every report decides who takes the credit for a sale. That decision comes from a piece of software following a rule somebody chose. Change the rule and the winner changes with it.
So treat attribution as a lens rather than a verdict. Ask whether a channel is likely to be doing what the report claims. Then look for a second signal, such as what people tell you on the phone.
The last click takes credit it did not earn
Somebody searches your company name, clicks the ad, and buys. The ad takes the sale. Something else sent them looking, though, and that something appears nowhere in the report.
This is why brand campaigns look weak on a spreadsheet, and why cutting them can hurt in a later month that nobody links back to the decision. In short, the numbers you can see are not the whole picture. Data driven marketing works best when it admits that.
A worked example of moving budget
Arithmetic makes a trade clearer than words do. The figures below are invented and round on purpose.
A worked example, with round numbers chosen to make the arithmetic easy rather than to describe a typical result. Two campaigns run for one month. Campaign A takes 600 of the budget and returns 30 enquiries, so each enquiry costs 20. Campaign B takes 400 and returns 10 enquiries, so each one costs 40. Now move 200 from B to A. If A holds its rate, that 200 buys 10 enquiries where it used to buy 5. Total spend has not moved, yet enquiries go from 40 to 45.
The sum is the easy part. The judgement is whether Campaign A can hold that rate at a bigger size, and it often cannot. Costs usually rise as you buy more, because the cheapest audience goes first.
So move budget in steps and check after each step. A large move made on one month of figures is a guess wearing a spreadsheet.
Data driven marketing when your numbers are thin
Small businesses hear that they should be data driven, then discover they have almost no data. That is normal. It is also workable.
Begin by counting three things you can count honestly. How many people enquired. Then how many of those were genuine, and how many actually bought. Three numbers kept by hand in a spreadsheet beat an analytics setup nobody trusts.
Then add one question to your enquiry form. How did you hear about us? It is imperfect, since people forget and guess. Still, it catches the word of mouth that no tracking tool will ever see.
Proper measurement comes next, once volume justifies the setup. Our guide to conversion tracking you can trust covers the parts that usually break first.
Making it a monthly habit
A quarterly review sounds sensible and rarely happens. Meanwhile a short monthly one tends to survive contact with a busy business.
Keep it to one page, in the same shape every time. What changed, why it changed, and what we are doing about it. Numbers appear as evidence for those three answers rather than as the content.
- Update the same four numbers, in the same order, on the same day.
- Write two sentences on what moved and why it might have moved.
- Name one change for next month, with the number you expect it to shift.
- Check last month's change before you commit to a new one.
Step four is the one everybody skips. Without it you pile up changes and never learn which of them worked. In practice this small loop is what data driven marketing looks like inside a real business.
When data driven marketing is the wrong answer
Sometimes the honest answer is that the numbers cannot settle this one. Pretending otherwise wastes money and time.
New channels carry no history, so the first months buy information rather than sales. Brand work pays back slowly and shows up later as searches for your name. Long sales cycles hide the result of today's spend for months.
In those cases, set a budget you can afford to lose. Agree how long you will give it. Write down what would count as a success before you start. That is still a decision built on evidence, and it is simply evidence you have not collected yet.
Deciding which channel deserves that trial is a separate question. Our guide to choosing marketing channels without spreading too thin works through it.
Why data driven marketing needs one change at a time
Two changes in the same month leave you with one result and no way to split it. Budget moves, a new landing page and a fresh headline all land together. Enquiries rise, and nobody can say which one did it.
So change one thing, then wait. The wait is usually longer than people expect. A slow week and a busy week look like a trend until several of them line up. Give a change at least as long as your normal gap between an enquiry and a sale.
Sometimes you cannot wait, since a season is arriving or a campaign is losing money now. Change both things in that case. Then write down that the two share one result, so nobody credits the wrong one later.
Ultimately data driven marketing is not sophisticated. It is a small set of honest numbers, read at the same time each month, by somebody willing to act on them.
Frequently asked questions
What is data driven marketing in plain English?
It is the habit of letting last month's results decide next month's spending. Nothing about it requires software. An owner with a notebook, four numbers and a fixed day in the month is doing data driven marketing, while a company with twelve dashboards and no decisions attached to them is not. The word data makes it sound technical. In practice it means you check what happened before you spend again.
How much data do I need before it is useful?
Enough that one extra enquiry would not change your answer. Here is a quick test. Work out the gap between the channel you want to back and the next one, then compare that gap with how much your months normally wobble. If the gap is smaller than the wobble, you have measured noise. Three or six months of totals often settle what one month cannot. So wait rather than move money on a difference that could reverse itself.
Which numbers belong on the front page of a report?
Only the ones you would act on, which is usually three or four. Cost per genuine enquiry and enquiry quality earn their place, since those two settle most spending questions. Impressions, sessions and follower counts belong further down, because nobody moves a budget over them. Show each number next to the last few months as well. A single figure with nothing beside it invites the reader to invent a trend.
Why does my report disagree with the ad platform?
Because the two count differently, and neither is lying. An ad platform usually credits itself when a sale follows a click inside the window it sets, while your own records may credit whatever came last. Time zones, refunds and blocked tracking widen the gap further. Pick one source as the number you act on and keep it. Then treat the other as a second opinion, useful for spotting a change in direction rather than for settling the total.
Is attribution software worth paying for?
Usually not until your volumes justify it. Attribution is a model rather than a fact, so a better model still produces an opinion about who earned the credit. A small business often gets more from one honest question on the enquiry form and a spreadsheet kept by hand. Once you spend across several channels every month, and the gap between them is worth real money, a proper tool starts to earn its cost.
How often should I review where the budget goes?
Decide monthly, on the same day, from the same one page. Check weekly only for faults, such as a broken form, a paused campaign or spend running away with itself. That weekly look is maintenance rather than a decision. If your trade has a season, compare this month against the same month last year as well as against last month, because a quiet month can be the season rather than the marketing.
When is data driven marketing the wrong approach?
When the numbers cannot settle the question yet. A new channel has no history, so early spend buys information before it buys sales. Brand work pays back slowly, and long sales cycles hide today's result for months. Data driven marketing still helps here, although you judge earlier signs instead. Watch enquiry quality, what callers say they saw, and whether your name gets searched more often. Agree the budget and the deadline in advance.
How do I start data driven marketing on a small budget?
Count three things by hand: enquiries received, enquiries that were genuine, and customers won. Add one question to your enquiry form asking how people heard about you. Those four inputs answer most of the questions a small business actually has. Skip the tools for now, because a dashboard nobody opens costs more attention than it returns. Once volume makes tracking worth setting up properly, add it and keep counting by hand for a month to check the two agree.
Should our agency write the monthly report?
They can, as long as you keep control of what the numbers mean. Agree the definition of an enquiry before the first report, or the figure will quietly include spam and wrong numbers. Ask for the same page every month, covering what changed, why it might have changed, and what they will do next. If a report never leads to a different decision, then data driven marketing has not started, whoever writes it.