On this page
The upgrade path most bars follow looks like progress. Paper, then a spreadsheet, then a reporting tool with charts, then a dashboard product with live data. Each step makes the same activity faster and prettier.
At no point does the activity change. Throughout, a person is looking at numbers and deciding what they mean. The tooling has become better at presenting the numbers and no better at the deciding, which is the part that was always the bottleneck.
What a spreadsheet actually does
A spreadsheet answers questions you already knew to ask. You decide the comparison, you build the view, and it computes. That is genuinely useful, and it has a specific limitation: it can never tell you about something you did not think to look at.
A dashboard is the same operation with the questions fixed in advance by whoever designed it. It is faster and less flexible. It still only surfaces what somebody anticipated.
For a bar this matters because the useful findings are usually specific and unanticipated. Comps clustering in one hour on one night. Variance concentrated in two products. A section that slows every Thursday. None of those are on a standard dashboard, because a standard dashboard is designed for the general case.
The three questions worth asking of any tool
Does it tell me something I did not ask?
If every insight requires the operator to have framed the query, the tool is a calculator. That is a legitimate product and it should be priced like one.
Does it arrive in time to matter?
A finding about last month can inform a policy. It cannot change an outcome. Most bar reporting is monthly because the accounting cycle is monthly, which is a reason about bookkeeping rather than about operations.
Does it say who does what?
This is the one almost nothing in the category does. A number is addressed to nobody. Until a finding becomes a specific action assigned to a specific role, the tool has moved work rather than removed it, and the work it moved is the hard part.
Why the category stopped where it did
Not incompetence. Measurement is tractable and interpretation is not.
Counting stock, reading a transaction feed and rendering a chart are well defined problems with correct answers. Deciding that a particular pattern means something, and that the something is worth a manager's attention tonight, requires a model of how a bar actually works. That is harder to build, harder to sell, and much harder to demonstrate on a sales call, because the output is a judgement rather than a screen.
So the category optimised the tractable half. The result is that a well equipped bar in 2026 can measure almost anything about itself and is no better at deciding what to do than it was, which is why buying more tools has stopped producing an improvement.
What the alternative has to demonstrate
Being sceptical of dashboards is easy. The replacement has an obligation to show three things, and it is fair to demand them.
It should surface a finding the operator did not ask for. It should do so inside the window where acting is still possible. And it should name who is doing what, then record whether it happened.
That last part is also the honest test of whether any of this works, because it is the only one that produces evidence. A tool that generates findings and never records outcomes cannot be evaluated, which is convenient for the vendor.
The comparisons go through the specific products one at a time, including what each does well. CoreTAP is the execution layer, and it connects to the point of sale you already run.