Excel is not the problem. We want to say that clearly at the start, because a lot of writing on this subject treats spreadsheets as something to be embarrassed about, and that is neither fair nor useful.
Excel is an extraordinary tool. It is flexible, everyone knows it, and it lets a competent person answer a question in twenty minutes that would take a week to specify to a developer. Most businesses should be using it for more things, not fewer.
The question is not whether Excel is good. It is whether your particular reporting process has crossed the line where Excel has stopped being the fastest way to get an answer and started being the reason answers take three days.
Here is how to tell.
Six signs you have crossed the line
None of these on their own is decisive. Three or more together usually is.
1. Preparing the report takes longer than reading it.
If someone spends two days assembling a pack that management reviews for forty minutes, the ratio is wrong. That is a signal regardless of anything else.
2. The same number is calculated differently by different people.
This is the most reliable indicator. When sales and finance present different revenue figures in the same meeting, the underlying problem is not arithmetic — it is that no agreed definition exists. Excel makes it easy for two correct calculations to disagree, because each file encodes its own assumptions invisibly.
3. Only one person can produce it.
If a specific colleague's leave delays the monthly reporting, the process is a personal skill rather than a business capability. This is a genuine operational risk that is usually only recognised when that person resigns.
4. The data is stale before it is read.
If the report describing March arrives on the fifteenth of April, it is a historical record rather than a management tool. You can record with it. You cannot correct with it.
5. You cannot answer follow-up questions.
The clearest symptom. Someone in the review asks "why is the North region down?" and the answer is "I will look into it and come back next month." A summary you cannot drill into is a summary you cannot act on.
6. Version confusion is routine.
Files named Sales_Report_Final_v3_REVISED.xlsx circulating by email, and genuine uncertainty about which one is current. This is not a naming discipline problem. It is a structural consequence of using files as a distribution mechanism.
Reasons to stay on Excel
Equally important, and less often said. Do not move if:
- Your reporting takes a couple of hours a month and nobody is complaining
- Reports change shape constantly, because you are still working out what to measure
- You have one or two data sources and they are already clean
- The people using the reports genuinely want to manipulate the data themselves
- Nobody has the appetite to agree on definitions, which is the real work
That last point deserves emphasis. A dashboard built without agreed definitions does not fix disagreement about the numbers. It automates it, and gives it an authoritative-looking interface. This fails more often than any technical problem does.
What takes the time
The most common misconception is that building a dashboard is mostly about building a dashboard. It is not. In a typical project the effort splits roughly like this:
Getting the data out (the largest part). Your ERP, CRM, ecommerce platform and spreadsheets each store things differently. Extraction has to be automated, scheduled and monitored, because a pipeline that silently stops is worse than no pipeline at all.
Cleaning and matching (larger than expected). The same customer exists as three records with slightly different names. Product codes changed in 2022 and history was not restated. Some entries have no date. None of this is visible in a spreadsheet where a human quietly fixes it each month. It becomes very visible when you automate.
Agreeing definitions (the part people skip). What counts as revenue — order date or invoice date? Does it include tax? Are inter-company transactions excluded? Every one of these must be decided, written down and agreed by the people who will later quote the number. This is a business exercise, not a technical one, and skipping it is the single most common cause of an abandoned dashboard.
Modelling. Structuring the data so questions can be asked efficiently, with consistent hierarchies for time, product, customer and location.
Building the visuals (the smallest part). Genuinely the easiest step once everything above is done.
If a proposal treats this as primarily a design exercise, be sceptical.
Designing for the person reading it
Dashboards get ignored when they are designed for the person who built them rather than for the person who reads them. Three principles help.
Different audiences need different screens. A managing director wants a handful of numbers with trend and a clear signal about what needs attention — readable in under a minute. An operations manager wants detail, exceptions and the ability to filter. Putting both on one screen serves neither.
Lead with exceptions, not totals. The useful question is rarely "what is our revenue." It is "what is unusual." A dashboard that opens with what is off-track gets used daily. One that opens with a grid of totals gets checked monthly out of duty.
Show the definition next to the number. If someone can hover over a metric and see exactly what it includes, arguments end quickly. This small feature does a disproportionate amount of work.
Refresh frequency: be honest
There is a strong pull towards real-time, and it is usually the wrong instinct.
Ask what you would do differently with hourly data versus daily. For most SME decisions — pricing, purchasing, staffing, credit — the honest answer is nothing. Daily refresh is dramatically cheaper to build and operate, and it is sufficient for the great majority of management reporting.
Real-time genuinely matters for operational screens: consignments in transit, orders awaiting dispatch, tickets breaching a deadline. Those are different from management reporting and it is worth being clear about which one you are building.
Do not throw Excel away
The best outcome is usually not replacing Excel. It is changing its role.
Today, Excel is probably doing everything: storing data, transforming it, calculating measures and presenting results. After a dashboard project, the pipeline and the model handle storage, transformation and calculation. Excel becomes what it is genuinely best at — ad hoc analysis on top of trustworthy, consistent data.
Every dashboard should export cleanly to Excel, and scheduled Excel or PDF reports should still go to people who prefer them. Insisting that everyone abandon spreadsheets is how you generate quiet resistance and end up with a parallel shadow reporting process, which is the worst of both worlds.
A sensible way to start
If several of the signs at the top apply to you, we would not start with a full project. We would start like this:
- Pick the one report that causes the most pain. Usually the monthly management pack or a weekly sales summary.
- Time it honestly. How many hours, from how many people, and how long after period end does it land.
- Run a definitions session. Get the relevant people to agree, in writing, what each measure includes. Do this before any tool is chosen. It costs a couple of days and it will tell you a great deal about whether the organisation is ready.
- Automate that one report end to end. Extraction, model, dashboard, scheduled distribution. Prove the pipeline on something narrow.
- Reconcile until it matches. Run the new output alongside the manual one until the numbers agree exactly. Credibility is established once, and lost permanently.
- Extend from there. With one working pipeline and an established pattern, each additional area is considerably cheaper than the first.
The short version
Move when producing the report costs more than the report is worth, when the numbers are disputed, when only one person can produce them, or when you cannot answer follow-up questions.
Stay on Excel when reporting is quick, the data is clean, and the shape of what you measure is still changing.
And whichever you choose, agree your definitions in writing. That single step improves reporting more than any tool will, and it is free.
