Data-Driven Marketing for Small and Mid-Size Companies: Start with Decisions, Not Dashboards

"Data-driven marketing" sounds like something for companies with analysts and expensive software. In practice, most small and mid-size companies already have more data than they use: in the cash register, the CRM, the ad accounts, website analytics, even in WhatsApp chats. The problem is rarely a lack of data. It is a lack of clear questions and of a routine for acting on the answers.

Start with decisions, not data

Write down the three to five marketing decisions you make again and again. For most companies the list looks something like this:

• Where should next month's budget go?
• Which products or services should we promote?
• Which customers should we bring back, and how?
• Do our discounts bring new customers, or do they only cut the margin on sales we would have made anyway?
• Which channels bring customers who stay and buy again?

For each decision, ask: what information would change my mind? That is the data you need. Everything else can wait.

The minimum data set

You do not need to track everything. You need a few sources that are connected and consistent.

Three rules make these sources usable:

1. One customer identifier. Choose one key that links a person across systems. In markets where many orders come through phone calls and messengers, the phone number often works better than an email address.
2. Consistent names. Use the same product, campaign and channel names everywhere. If one person writes "Insta" and another "IG paid", your reports will never add up.
3. Tag every link. Add UTM parameters to the links in ads, posts, newsletters and profile bios, so web analytics can tell you where visitors came from.

Prioritise first-party data: information customers give you directly through purchases, sign-ups and conversations. Third-party tracking has become less reliable. Browsers such as Safari and Firefox restrict third-party cookies by default, and iPhone apps must ask users for permission before tracking them across other companies' apps and websites. Your own data is more stable, and you control its quality.

Collect only what you need, ask for consent where required, store data securely and follow your country's personal data law. Trust is part of the data strategy.

Simple analysis that answers most questions

You do not need advanced statistics to start. These six analyses cover most everyday decisions and can all be done in a spreadsheet:

1. Trends over time. Weekly and monthly figures, compared with the same period last year to separate seasonality from real change.
2. Segments. New versus returning customers, by channel, region, product and price range. Averages hide the story; segments show it.
3. Repeat purchase. Of the customers who first bought in a given month, how many came back within 30, 60 or 90 days? This tells you more about the long-term health of the business than any campaign report.
4. RFM. Score customers by recency, frequency and monetary value. It is the simplest way to see who your best customers are and who is about to leave.
5. Unit economics. Compare the cost of acquiring a customer with the gross margin that customer brings over time. If you lose money on every new customer and they never return, more budget only makes it worse.
6. Simple tests. Change one thing at a time — an offer, a headline, a channel — and compare the result with a control group, period or region where nothing changed.

Two warnings. First, do not trust platform reports blindly. Each ad platform measures its own contribution in its own way, and the conversions they report can add up to more than your actual sales. Check them against your own sales records. Second, correlation is not causation. Sales went up during the campaign — but was it the campaign, the season or a competitor running out of stock?

A dashboard people actually use

A dashboard is useful only if someone makes a decision with it. Keep it to one page and around ten metrics, in three layers:

• Business results — revenue, gross margin, number of new and returning customers.
• Marketing efficiency — customer acquisition cost, cost per lead, return on ad spend.
• Leading indicators — website traffic from target channels, enquiries, response time, conversion from enquiry to sale.

For each metric, write down a definition, a data source, an owner and a target. A spreadsheet connected to a free visualisation tool such as Looker Studio is enough to begin with; what matters is that the numbers are updated on a fixed schedule. Leave out vanity metrics such as follower counts or impressions with no link to sales.

From numbers to decisions

Hold a 30-minute review every week around the same three questions: what changed, why do we think it changed, and what will we do about it? Write each decision down with the reason and a date to check the result. After a few months this log becomes one of your most valuable marketing assets: a record of what actually works for your business.

AI assistants can help with the routine work: writing spreadsheet formulas, summarising trends, flagging unusual changes. Check their output against the source numbers, and never paste personal customer data into tools that are not approved for it.

Common mistakes

• Collecting everything and analysing nothing.
• Treating platform dashboards as the truth.
• Launching campaigns without a baseline, so nobody can say whether they worked.
• Reporting vanity metrics because they always go up.
• Ignoring qualitative data — reviews, sales notes, support chats — that explains the numbers.

Where to start

Choose one decision you make every month. Define the three numbers that should inform it, find where each one lives, and build a single sheet you update every Monday. In my experience, one decision made consistently on evidence changes a team's habits faster than any large analytics project.