Read a sample: 1. Start With the Decision, Not the Dashboard
On Monday morning, a small-business owner opens a marketing dashboard and sees a welcome sight: the latest campaign reached more people than usual. Comments are up. A video has been shared. The numbers look like evidence that the business is getting somewhere.
Then the owner checks the sales record. Revenue is essentially unchanged.
Both observations can be true. More people can notice a business without more people buying from it. Some may live outside its service area. Some may enjoy the content but have no need for the product. Others may be interested but unable to find the next step, or they may intend to buy later. The dashboard has not necessarily misled anyone. It has answered a narrower question than the owner wanted to ask.
That gap is where useful marketing measurement begins.
The purpose of measurement is not to collect the largest possible set of numbers. It is to make a better decision: whether to continue an activity, change its message, improve a handoff, investigate a weak point, or stop spending time on something that is not serving the business. A number becomes useful when you know what question it helps answer and what you might do differently after seeing it.
For a small business, this approach is especially valuable. You may be the owner, the marketer, the sales team, and the person answering customer questions. Even if you have help, you probably cannot investigate every possible explanation for every change. A focused question protects your attention as well as your budget.
Marketing activity is what you do. A business outcome is the result you hope that activity will help produce.
Publishing an article, sending an email, attending an event, and running an advertisement are activities. More qualified inquiries, more purchases, a higher proportion of customers returning, and stronger contribution after relevant costs are possible outcomes. Some outcomes sit between marketing and the final sale. An inquiry is not revenue, for example, but it may be an important step toward revenue for a service business.
The distinction matters because completing an activity does not establish that it worked. A campaign can launch on schedule, generate attention, and still fail to move the business toward its objective. Equally, an activity can have a worthwhile purpose before a sale occurs. A new business may need more relevant people to learn it exists; a consultant with plenty of inquiries may need to improve follow-up rather than increase visibility.
Consider these different questions:
- Activity: Did we send the email?
- Attention: Did people open it, click it, or visit the page?
- Customer response: Did suitable prospective customers inquire or take another meaningful next step?
- Business outcome: Did purchases, repeat business, or contribution improve?
Each question can be legitimate. Trouble starts when an answer to one is presented as proof of another. Sending an email does not prove that anyone read it. A click does not prove that the visitor was ready to buy. A purchase does not, by itself, prove which marketing activity caused it.
That does not make marketing measurement hopeless. It makes the task more precise. You are looking for evidence that fits the decision, while being honest about what the evidence cannot establish.
Imagine a hypothetical neighborhood bakery called Cedar Street Bakery. Its owner, Lena, wants to increase orders for custom celebration cakes. She has room in the production schedule and believes the cakes earn enough, after their direct costs, to make additional suitable orders worthwhile. Before committing more staff time to promotion, she wants to know whether her recent social media effort is helping generate those orders.
For this hypothetical example, Lena posted cake-decorating videos during a four-week period. Her platform reported 24,000 views, compared with 8,000 views during the preceding four weeks. The increase is easy to calculate: 24,000 minus 8,000 is 16,000 additional views. But a view is an observation recorded by the platform, not necessarily a unique local person or a prospective customer. Someone may have watched more than once, and people outside the bakery’s delivery area may be included.
Lena also checks the bakery’s order records. In the earlier four weeks, customers placed 20 custom cake orders. During the video period, they placed 20 custom cake orders again. On those records, orders were flat.
The two measures describe different things. Views increased; recorded orders did not. Neither observation erases the other. The videos may have introduced the bakery to new people, entertained existing customers, or contributed to orders that would otherwise have gone elsewhere. They may also have reached many people who were unlikely to order. The available figures do not tell Lena which explanation is right.
The comparison has other limits. The two periods may have differed in holidays, local events, customer demand, or the amount of time people needed to plan a celebration. A customer might see a video in the second period and order after it ends. With only these two short periods, Lena should be cautious about treating the first as a dependable baseline or claiming that the videos caused—or failed to cause—a change in sales.
Still, she has enough information to ask a better question. Instead of asking, “How do I get even more views?” she can ask, “Are these videos bringing suitable cake customers closer to placing an order?”
That question leads her to examine what happens between viewing and purchasing. In the hypothetical second…