Loading
CoreTAP connects to your Toast POS and turns every transaction into a named, assigned, tracked action, not another dashboard.
Connects to the POS you already run. No register swap.
Void rate on Bar 2 is running well above the team on the same daypart.
Every tool in the category stops at the second step. CoreTAP carries on to the last three.
Nearly 2x
Revenue per two weeks rose from $45k to $50k up to $80k to $90k, at a Houston hotel bar, after somebody acted on what the point of sale had been saying.
One venue, over a single reporting period. The registry figure is a range on both sides, so the headline is the multiple rather than a single pair of numbers it would have to round to. Read what changed. Client reported, from an anonymised venue, and described by type rather than named at their request.
Every module finds something, works out what it means, and hands the next move to a named role with a date. The names are the product’s own, so nothing here has to be translated on a demo.
You find out during the shift instead of on Monday, which is the difference between changing an outcome and explaining one.
The current service while it is still running. This is the view that makes the difference between a finding and a decision, because everything else on this list is only useful if somebody sees it in time.
One thing gets fixed each day and stays fixed, instead of eleven things being noticed and none being followed up.
A single improvement for the day rather than a wall of metrics. One action a day, checked the next, is the cadence that survives a week where a supplier fails and someone resigns.
The leak gets a name, a night and an owner, so it stops being a variance figure and starts being a fix.
Where the money leaves without a sale attached. The largest module in the product, because this is where most recoverable margin sits.
What is in itStaffing matches how the room actually trades, so you stop paying for the hours that do not sell and stop being short in the ones that do.
When the venue makes money and when it does not. A room does not trade at one speed, and the hours that carry a week are rarely the ones people assume.
What is in itCoaching lands on a specific shift with a number attached, instead of a general conversation nobody can act on.
Performance by person and by grade, so a conversation is about a specific shift rather than a general impression.
What is in itThe serves that stopped earning get changed or repriced while it still matters, rather than at the next menu print.
What sells, what does not, and what the mix says about both. Menu Ideas proposes changes from the trading pattern rather than from a template.
You can tell whether anything you decided actually happened, which is the only way to know if any of the rest of this worked.
Every finding in every module ends here, assigned to a role with a date, and closed with an outcome. A report is addressed to nobody. An action is somebody's job, and the record of whether it happened is the only way to tell if any of this worked.
You see returning spend fall before the revenue line does, which is the earliest warning this business gets.
Return behaviour and spend over time, from the transaction record rather than from a survey.
A new starter reaches the standard in weeks instead of months, and the good ones have a reason to stay.
A flagged habit becomes specific coaching against the shift it came from, rather than a general module about bartending.
Prime cost is a number you can still act on, rather than a verdict delivered six weeks after the money went.
Invoices, prime cost, variance and margin, on the same clock as the shift. It is a product in its own right and it is also a module inside CoreTAP.
Two shifts carry it. A monthly figure would average this away and leave you looking for a policy fix to a rota problem.
One hour, not a generous team. Something went wrong at eight and the fix was to give drinks away. That has a cause somebody on the shift remembers.
Bar 2 stopped selling at nine. At venue level the night looks soft. At section level it has an address and a staffing decision attached.
Same covers, lower spend. The mix moved to the cheaper serve, which is a selling problem rather than a footfall problem.
Illustrative figures. The shapes are the findings this page describes; the numbers are not from any one venue.
No hardware on the bar, no second tablet, and no change to how a bartender rings a round. CoreTAP reads transaction level data from the system in place. What integration actually means sets out the four questions worth asking any vendor, including us.
Named in type rather than shown as logos, because we do not licence these marks and a wall of borrowed logos tells you nothing about what actually moves. Ask us what data we read from yours, and how late it arrives.
An owner with one venue can stand in the room. They know which bartender is heavy-handed, which section drags on a Friday, and roughly what the walk-in should look like on Sunday morning. That knowledge does not survive the second location, and it definitely does not survive the fourth.
What usually replaces it is a weekly P&L and a phone call. Both arrive too late to change anything, and neither can tell you whether a bad week was demand or execution. That distinction is the entire job. Demand you plan around. Execution you fix.
CoreTAP reads from your existing point of sale. No register swap, no new terminals, no retraining the floor on hardware. Historical data is available as soon as the connection is live, so the first read is usually same day rather than the start of a 90 day baseline.
Every transaction gets attributed. Not just what was sold, but who rang it, on which shift, at what hour, with which modifiers, and against what it should have been. This is where the two categories separate: a rough night that was genuinely slow, and a rough night that was slow because of how it was run.
Findings route to the person who can do something about them, at a time when doing something is still possible. A pour variance flagged Thursday is four kegs cheaper to fix than the same variance found in next month's reconciliation.
That timing is the whole argument for real time restaurant analytics over monthly reporting. A monthly report is an autopsy. It tells you accurately what killed the margin, several weeks after anything could be done about it. CoreTAP is a restaurant intelligence platform rather than a reporting tool because the output is meant to change the shift it describes.
Sales by item, category, daypart and hour, with the pour cost carried through so contribution is visible rather than inferred. Most menus contain two or three items that look like winners on volume and lose money on margin. They are usually the ones nobody questions.
Sales per labor hour, void rate, comp rate, discount usage and average check, per person, benchmarked against the rest of the team on comparable shifts. The point is not surveillance. The point is that venue-wide averages conceal the two people at either end of the distribution, and those two are where both your problem and your next promotion live.
Draft beer monitoring compares what physically left the keg against what was rung. A few ounces of overpour is invisible on any single drink and substantial across a quarter. It is the most consistently underestimated loss in a bar, because nothing about it looks like theft. It looks like a generous bartender.
Voids after close, comps with no matching approval, discount codes clustering on one operator, refunds at odd hours. Individually explainable. In a pattern, not. CoreTAP surfaces the pattern rather than the incident, which is the difference between an accusation and a conversation.
Where labor hours actually went against where the demand actually was. Overstaffing is expensive and quiet. Understaffing is expensive and loud, and it shows up later as turnover and bad reviews rather than as a labor line.
Draft is worth calling out separately because it is where the largest recoverable losses usually sit and where operators have the least visibility. A keg holds a known volume. A pour has a target volume. The gap between what the keg gave up and what the register recorded is your real variance, and until you measure it you are guessing.
The reasons that gap opens are mundane. Foam on a line that needs balancing. A glass filled generously because the guest is a regular. A tap left running during a rush. A keg changed mid-service with half an inch unaccounted for. None of that is theft, and none of it shows up as an event anywhere in your system. It shows up as a cost percentage that is two points worse than it should be, which then gets blamed on distributor pricing.
Once the variance is attributed by tap, by shift and by person, the conversation changes from "our pour cost is high" to "this line has been running heavy since the eleventh, and it is worth this much a week." One of those is actionable.
There is no ninety day baseline period, because the data already exists. The POS has been recording every transaction for as long as you have been open. Connecting CoreTAP makes that history readable rather than starting a new collection window.
In practice the first read usually surfaces three or four things the operator half suspected and could never prove, plus one they did not see coming. The half-suspected items are typically a shift that underperforms, a comp rate that has drifted, and a menu item priced against costs that changed. The surprise is usually staffing shape: labor hours that were distributed evenly across a night when demand was not evenly distributed at all.
Two weeks of trading is normally enough to separate a pattern from noise. One bad Saturday is a bad Saturday. The same variance on the same shift for six consecutive weeks is a process, and processes can be changed.
Delta Kinetics was founded by a researcher with a doctorate, and CoreTAP is grounded in analysis of how bars and restaurants actually operate. We published a six volume research program on it: 13,063 guest reviews across 41 venues in four Texas markets, every review classified twice by separate models, checked against state alcohol tax filings. That matters for a specific and unglamorous reason: benchmarks are only useful if the comparison set is real. When CoreTAP flags something as abnormal, it is abnormal against measured venues, not against a round number somebody chose because it sounded reasonable. The full method is public.
CoreTAP is the execution layer. It pairs with multi-location rollups when you are comparing venues rather than a venue against target, and with LearnTAP when the finding is a training problem rather than a staffing one. Most operational issues turn out to be one or the other, and telling them apart early is most of the value.
If you want the category-level view rather than the product detail, start with restaurant analytics. If your problem is specifically behind the bar, bar management software is the closer fit.
CoreTAP is $999 per month, or $799 per month billed annually. Full pricing and what is included.
Bring a two week window from your venue. We will show you where the variance is before you commit to anything.