A meeting without numbers is a trial with four lawyers and no evidence. Everyone talks, everyone has an explanation, and the winner is whoever is best at talking their way out of it. My rule for data-driven decision-making is simple: numbers first, context second.
![\"Marketing](\"https://soymarketingultra.com/wp-content/themes/astra-child/assets/brand/despiram-sorprendido.png\")
TL;DR: The No-Nonsense Summary
Data-driven decision-making means using relevant measurements to describe a situation, test hypotheses and choose an action. Showing a beautiful dashboard, pointing at a green number and calling the conversation over is something else.
When I review an operation, I want to see workloads, time tracking, efficiency, committed dates and progress. Then we can talk about the client, scope changes, unexpected complexity or Mercury being in retrograde. Context matters, but it has to explain an observed reality.
Monitoring data makes it possible to control resources, activities, outputs and outcomes, spot deviations and guide corrective action. The UK Government's Magenta Book says as much. That is the point: putting a shared baseline on the table.
Without that baseline, your opinion is worth exactly the same as anyone else's. It may be brilliant. It may come from twenty years of experience. It may also be a story you have made up to explain why something went wrong.
A useful operational report must connect what was promised, the resources used, the progress achieved and the outcome. A report full of hours may only be measuring utilisation. Focus on completed tasks and you end up rewarding whoever breaks them into the smallest pieces. And if you wait for invoicing, you are already too late.
\n![\"Mascot](\"https://soymarketingultra.com/wp-content/uploads/data-driven-decision-making-numbers-first-diagrama.webp\")
I would start with five questions: what was committed, by when, how many hours were planned, how many have been used and what progress or outcome exists? That lets you locate the discrepancy. The conversation stops being “we are in trouble” and becomes “we have used 80% of the hours with 45% of the work complete.”
There. Now we have something to discuss.
Everyone on the team needs to log their work, and do it properly. Because you can have up-to-date records and still have a pile of rubbish. Writing “miscellaneous tasks” for six hours counts as time tracking. Technically.
This requires discipline and a shared way of planning. I would use Scrum for marketing team organisation to create that rhythm, provided the method helps make work visible rather than becoming another religious ceremony.
No. Numbers eliminate some pointless arguments and improve the questions. You still need judgement to distinguish a cause from a coincidence, an urgent issue from a symptom, and a valuable outcome from a metric that merely looks good.
A study of more than 18,000 US manufacturing plants found that late adopters of data-intensive practices had, on average, 3% more value added than non-adopters. The Brynjolfsson and McElheran paper itself notes that its design was consistent with a causal relationship, but did not prove one.
Performance was not uniform either. It depended on complementary factors such as available technology, worker skills, company type and decision-making structure. Data helps. The organisation using it still has to work.
In my experience, this is where the most common trap appears: confusing activity with value. More logged hours do not mean more productivity. More completed tasks do not mean better work. More campaigns launched do not mean more business. Reward the wrong counter and the team will learn to inflate it.
Context should come second because, without a shared reference point, it can easily become an excuse. Once you have identified the deviation, it is time to understand what caused it, how far it reaches and what might break if you correct it.
The official DORA metrics guidance warns against two particularly tempting mistakes: comparing teams or systems in different contexts, and reducing a complex system to a single indicator. It also reminds us that turning a metric into a target increases the risk of gaming it.
Want to improve speed? You can do it by shipping work too early and wrecking quality. Want to increase utilisation? You can fill every gap in the calendar and destroy your capacity to react. Want more sales? You can accept bad projects and celebrate the revenue until the trouble arrives.
A decision based on opinions can break something else. So can a decision based on a distorted metric.
The OECD argues that evidence must fit the question and local context, consider alternatives, positive and negative effects, limitations and uncertainty. It also keeps human judgement in major decisions. That seems to me exactly the right order.
An Excel spreadsheet is not right by divine illumination. But an operational conversation needs solid ground. Bring me a detailed report of work completed, hours spent, planning and progress, and many discussions get shorter. The report does not make the decision. It makes the disagreement visible.
\n![\"The](\"https://soymarketingultra.com/wp-content/uploads/toma-de-decisiones-con-datos-primero-n-meros-luego-cont-ilustracion.webp\")
Does the data fail to measure what we need? Change it. Teams working in different contexts should not be casually compared. And if quality, impact or client satisfaction are missing, they need to be added. Measuring well also means acknowledging what we do not know.
Numbers first, context second. The conversation is still there, but it stops being a competition of narratives. That is the kind of data-driven decision-making I stand for: enough figures to look at the same problem, and enough judgement not to follow them like sheep.
", "markdown_url": "https://soymarketingultra.com/en/data-driven-decision-making-numbers-first.md" }