There are four main ways to measure draft beer kegs: manual estimation , weighing kegs on a keg scale, torque-based estimation with a Keg Check device, and real-time flow monitoring. Estimation is the fastest and the least accurate. A keg scale works on every keg size and is what Bar-i uses most often. Keg Check is quick but fits only two keg formats. Flow monitoring shows you what happens between counts, but it will not tell you whether what you poured matches what you sold. Which one fits depends on how many kegs you run and how much volume each one does.
You walk into the cooler on a Monday morning. There are eleven kegs down there, all of them stainless steel, all of them opaque, and not one of them will tell you how much beer is left inside. So you do what almost everyone does — you tilt one, feel the weight shift, and write down "0.5"
Then you write down "0.5" for the next one too, because it felt about the same.
That number goes into a spreadsheet. That spreadsheet becomes your beverage cost. And by Friday, someone is asking why draft beer looks expensive again this month, and the honest answer is that nobody actually knows — because the measurement at the front of the process was a guess.
Draft beer is the hardest category in the bar to measure well. Liquor bottles are glass; you can see the level. Wine comes in units you can count. A keg is a sealed metal cylinder that weighs roughly 155 pounds when it's full, sits on the floor of a walk-in, and gives away nothing.
That is exactly why it's worth understanding your options before you pick one.
With liquor bottles, experienced counters use "tenthing" — estimating to the nearest tenth of a bottle. With kegs, tenthing is usually impractical. You cannot see through the steel, and you are not going to feel the difference between 60% and 70% full in a 160-pound object.
So in practice, bars estimate kegs on a five-point scale: 0, 0.25, 0.5, 0.75, or 1.0.
Where it works: It's free, it's fast, and it requires no equipment. For a small draft list where each keg turns over quickly, a rough number may be all the operator needs.
Where it breaks down: It is the least precise method available, and it leaves no data trail. Two different managers will produce two different numbers for the same keg on the same night. There is nothing to audit, nothing to compare, and no way to tell whether this week's variance is real or just a different person's judgement. If hand estimation is your only option for now, at least standardise it — we've written before about the different ways to perform inventory on a beer keg and what each one costs you in accuracy.
Bar-i Pro Tip
Consistency beats precision. A slightly imperfect method applied the same way every single week gives you trend data you can act on. A precise method applied inconsistently gives you noise that looks like data.
Put the keg on a scale and read the weight. Subtract the empty keg weight, and you have the beer remaining.
This is the method Bar-i uses most frequently across our client base, and there are good reasons for it. A keg scale works on any keg size — half barrels, quarter barrels, sixtels, euro kegs, pony kegs. A keg scale displays weight in 0.1 lb increments, so the reading itself is not the weak link in your count. And unlike an estimate, a weight is a fact: the same keg weighed by two different people produces the same number.
A few practical notes. Kegs are heavy, so you need a scale built for the load, not a kitchen scale. Keg scales also need to be handled with a bit of care — they're precision instruments, and rough treatment shortens their life. Weighing removes the guesswork, but it doesn't remove the discipline — most of the accuracy still comes from counting the same way, in the same order, every single time.
Bar-i sells a tested keg scale through the Bar-i Store for $120 shipped; operators create a free Bar-i account to order one. We review the recommended model periodically based on pricing, availability, and field testing, so if you want a current recommendation, just ask us — we'd rather point you at the right scale than have you buy the wrong one.
Bar-i Pro Tip
Count your kegs at the same point in the cycle each time — the same day relative to your deliveries and your line-cleaning schedule. Otherwise, cleaning waste lands in one week and disappears from the next, and your yield swings for reasons that have nothing to do with your bar.
Keg Check USA takes a different approach. Instead of lifting the whole keg onto a scale, you balance the keg on its bottom corner and the device measures the twisting force — the torque — required to hold it there. More beer inside means more resistance, and the device converts that into an estimate of what remains.
It's genuinely fast once a counter has the technique down, and it's a meaningful step up from shaking a keg and guessing.
The limitation is keg size. Keg Check works on two formats: the standard 15.5-gallon US half barrel and the 13.2-gallon "euro keg" used by some US breweries. It does not work on quarter barrels, sixtels, or other sizes. It also requires some user skill and, ideally, a hard flat surface — if your kegs sit on shelving, they'll need to come down first.
For a high-volume venue running mostly half barrels, that's not much of a constraint. For a craft-focused bar with a long list of sixtels, it rules the method out.
4. Real-Time Monitoring and Flow Meters
Systems such as Kegtron, KegSafe, iDraft, BevChek, and BeerSaver install on the draft line and record beer as it flows through. They work on any keg size, and the appeal is obvious: instead of finding out on Sunday night, you can see pour activity as it happens.
The real benefit here is visibility between counts. If you want to know what happened on Saturday at 11pm — not just what the week totalled — a flow meter is the only method on this list that can tell you. This is the newest category of the four, and the technology for monitoring draft beer inventory in real time has moved quickly in the last few years.
Two things to keep in mind. First, vendors publish high accuracy figures, but those numbers are difficult to verify independently, so we'd encourage operators to evaluate these systems on the visibility they provide rather than on a published accuracy percentage. Second — and this is the bigger one — pour data on its own does not reconcile against your POS sales. Knowing 412 ounces left the tap doesn't tell you how many of those ounces were rung in. Someone still has to connect the two.
| Method | Works on | Precision | Data trail | Between-count visibility | Best fit |
|---|---|---|---|---|---|
| Manual estimation (five-point) | Any keg | Lowest | None | No | Very short draft list, high volume per keg |
| Keg scale | Any keg size | Reads in 0.1 lb increments | Yes | No | Most bars, including long and mixed draft lists |
| Keg Check USA | 15.5-gal half barrel and 13.2-gal euro keg only | Moderate | Yes | No | High-volume venues running mostly half barrels |
| Real-time flow meters | Any keg | Vendor-dependent | Yes | Yes | Complex or high-volume draft programs wanting live pour data |
Bar-i Pro Tip
Whichever row of that table you land on, the measurement only earns its keep once it meets your POS data. A keg weight that never gets compared against sales is a number without a job.
There is no universal answer, and anyone who tells you there is one is selling something.
The decision comes down to two separate volume factors that people tend to collapse into one:
A short draft list doing high volume per keg can be measured simply. A long draft list where each keg turns slowly needs a more precise method. And the industry has been moving toward longer draft lists for years, which means more bars are drifting into the second category without noticing.
Beyond volume, weigh your keg sizes (mixed sizes rule out Keg Check), your line setup and length, your storage layout (can you get a scale under a keg on that shelf?), and how much visibility between counts you actually need.
Everything above is about answering one question: how much beer is in the keg right now?
That's a useful question. It is not the question that makes you money.
The question that makes you money is: of the beer we poured last week, how much did we actually sell?
Measuring the keg is step one. The business value comes from connecting keg usage to your purchases, your recipes, and your POS sales data — so you can put what was poured next to what was rung in and look at the gap.
Sample live Bar-i report showing sold vs. poured down to the serving.
Bar-i Pro Tip
Check what your recipes assume about glass fill before you go hunting for a variance. A nominal 16 oz serving usually holds around 14.5 of beer when poured with a proper amount of head. If the recipe says 16 and correctly poured the serving is 14.5oz, your numbers will never reconcile no matter how well you measure the keg.
Draft beer makes that gap unusually hard to read, because some draft loss is genuinely unavoidable. Foam. Beer left in the lines. Beer left in the keg. Line cleaning. Startup waste before the system comes down to temperature. All of it is real, all of it is normal, and knowing what a normal draft beer waste level actually looks like is what lets you tell ordinary loss from the kind you should act on. Together it creates a smoke screen — a fog of legitimate loss that human-created loss can hide inside.
You cannot separate the two by counting harder. You separate them by measuring consistently, comparing poured against sold, and reviewing yield item by item rather than as one draft category number. High-volume beers tend to show strong yield. Low-volume kegs are usually where the damage is, and a single category-level number will never show you that. Another nuance is that sixtels and pony kegs can tend to drag your keg yield down due to additional changes they necessitate.
Bar-i Pro Tip
Watch your slowest movers first. The kegs that sit longest are where yield problems hide, and they are exactly the ones a single category-level draft number will never surface.
Across Bar-i's audit experience — 50,000+ audits coast to coast — the average bar pours materially more than it records as sold. Bar-i's own benchmark puts that gap at around 15%, and draft beer frequently runs worse than the overall average.
Want to see what that comparison actually looks like? Book a free Bar-i demo and we'll walk you through a real draft beer yield report — broken down by item, not by category.
Weighing the keg on a dedicated keg scale is the most broadly accurate method. It works on every keg size and reads in 0.1 lb increments. Real-time flow meters can also produce detailed pour data, but they measure what left the tap rather than what remains in the keg, and vendor accuracy figures are difficult to verify independently.
A full half-barrel keg weighs roughly 155 pounds. That weight is the main reason manual "tenthing" doesn't work on kegs the way it does on liquor bottles — and the reason a keg scale needs to be rated for the load.
Yes. Manual estimation on a five-point scale (empty, quarter, half, three-quarters, full) requires no equipment, and torque-based devices such as Keg Check USA estimate the remaining volume without lifting the keg onto a scale. Both trade some accuracy for speed and convenience.
No. Keg Check USA works on the 15.5-gallon US half barrel and the 13.2-gallon euro keg. It does not work on quarter barrels, sixtels, or other keg sizes. Bars running mixed keg sizes generally need a scale instead.
It depends on what you need. Flow meters are the only method that shows pour activity between counts, which is valuable for complex or high-volume draft programs. But pour data does not reconcile itself against POS sales — you still need a process that compares what was poured against what was sold. The tool creates the data; it doesn't interpret it.
A yield above 95% — meaning losses under 5% of total usage — is a useful operating target where the draft system and product mix make that realistic. What matters more than hitting a specific number is measuring yield consistently, by item, so you can see which lines are underperforming and why.
Bar-i uses keg scales as its primary method. Accountability data requires two counts — a starting count and an ending count. Bar-i calculates usage between those two counts and compares it against your POS sales data to show draft beer yield by item. POS integration is included in the Bar-i Complete and Pro subscription plans at no additional charge.
There is no single best way to measure every draft program. Rough estimates may be sufficient for a small, high-volume draft list. Bars with more kegs, lower volume per keg, or a need for between-count visibility may benefit from scales, Keg Check, or real-time monitoring.
But whatever tool you choose, the measurement only becomes useful when it's connected to your purchases and your POS sales — so you can see what was sold versus what was poured.
That's the number that changes decisions. Everything else is just dispense data.
See it on your own numbers. Book a free Bar-i demo and we'll run it against your draft list. Bar-i counts inventory fast — 500 items an hour with barcode scanning and Bluetooth weighing — and hands you a draft beer yield report broken down item by item, with a specialist who walks you through where the gap is coming from and what to do about it. No POS reconfiguration required to start; we work with your setup as it stands.
Prefer to watch first? Start with How it works: Counting and How it works: Down to the Serving Results in the Bar-i video library.
Next in this series: Periodic Keg Counts vs Real-Time Draft Beer Tracking: What Each Actually Tells You