Bar Operational Intelligence: A Definition and a Playbook

We should declare the obvious first. This company sells a product in this space, so a piece defining the category is not a neutral act. The way to make it useful anyway is to write a definition specific enough to exclude things, including things we sell, and to give the playbook away.

What is bar operational intelligence?

Operational intelligence is the practice of turning an operational data stream into a specific action, assigned to a specific role, inside the window where the action can still change the outcome, and recording what happened next.

Four clauses, and each one excludes something.

A specific action, not a number

This excludes most reporting. A variance figure is a measurement. Put the jigger back on the well gin tonight is an action. The step between them is the entire discipline, and it is the step almost every product leaves to the operator.

Assigned to a specific role

This excludes dashboards, which are addressed to whoever opens them. A finding with no owner produces agreement rather than change.

Inside the window

This excludes month end analysis, however good. A finding about a period that has closed can inform a policy and cannot change a result.

Recording what happened next

This excludes alerting products, and it is the clause most often missing. Without an outcome record there is no way to tell whether any of it worked, which is convenient for a vendor and useless for an operator.

How does it differ from reporting and business intelligence?

Reporting describes what happened, accurately, after the fact. It is an accounting function and it is necessary.

Business intelligence lets you interrogate what happened. It answers questions you thought to ask, which means it cannot surface what you did not think to look for.

Analytics finds patterns. It stops at the pattern, which is one step short of a decision and two short of an action.

Operational intelligence, as defined above, continues to the action, the owner and the outcome. That is the only distinction being claimed, and it is a distinction about where a process ends rather than about how clever anything is.

The playbook, without software

All of this is doable by hand. It is worth doing by hand first, because a venue that has run it manually knows what it wants from a tool.

1. Pick one stream

Transactions from the point of sale. Not inventory counts, not the profit and loss, not a survey. Transactions, because they are timestamped, granular and already yours.

2. Compare like with like

This Friday against the last four Fridays, at the level of product, hour, section and comp. Not against budget, and not against last month.

3. Find the largest explicable gap

Explicable is the operative word. The biggest gap may be weather. You want the biggest one with a plausible internal cause, which is usually smaller and always more useful.

4. Ask the people who were there

Within two days, while they still remember. They almost always know, and they have almost never been asked, because the number normally arrives too late for the question to make sense.

5. One action, one role, one week

Specific enough that next week it is obviously either done or not done. Assigned to a named role rather than to the team.

6. Check it first, next week

Before looking at anything new. Skipping this converts the whole practice into a weekly noticing exercise, and the team learns that nothing follows from it.

Where the honest limit is

That playbook takes about thirty minutes a week plus the assembly. The assembly is the problem: getting seven days of transactions into a shape where shifts are comparable is most of the work, it has to be redone every week, and it is the reason the routine dies around week four.

That is the entire case for tooling in this category. Not that software notices what a competent operator would miss, but that it removes the step that kills the habit. An operator with the discipline and the hours does not need it, and should be told so.

CoreTAP is our implementation of the definition above. The comparisons assess the other products in the category against the same four clauses, including where they are stronger than us.