Back to blog
Inventory Management

How to Use ABC Analysis to Plan Cycle Counts

October 2, 2026

A cycle count program lives or dies on one decision: which items do you count, and how often? Count everything equally and your team spends most of its time on slow, low-risk stock. Count only when something looks wrong and you are back to guessing. ABC analysis gives you a simple, defensible way to aim your counting effort at the items that matter most.

This guide walks through the idea behind ABC classes, how to rank your SKUs, how to set count frequencies, and how to turn all of it into a weekly schedule your team can keep.

The idea behind ABC classes

In most inventories, a small share of SKUs accounts for a large share of the value or the movement. A handful of key ingredients or best-selling finished goods carry the business, while a long tail of slow items sits on the shelf and rarely changes.

You will often hear this described as the 80/20 rule: roughly 20 percent of items drive roughly 80 percent of the value. Treat that as a rule of thumb, not a measurement. Your own split may be very different, and the only way to know is to rank your own data.

ABC analysis sorts SKUs into three groups based on that ranking:

  • A items: the small group at the top that drives the most value, movement, or risk. These get counted most often.
  • B items: the middle group. Important, but less so. These get counted at a moderate pace.
  • C items: the long tail. Many SKUs, each with little impact on its own. These get counted least often.

The point is not precision. The point is to stop treating a pallet of slow-moving spare cartons the same as the ingredient that goes into every batch you make.

Choose what you rank by

Before you can draw class lines, you need to decide what “important” means for your operation. There are three common bases, and each one answers a slightly different question.

  • Value: annual usage multiplied by unit cost. This ranks items by how much money flows through them. It is the classic approach and works well when financial accuracy is the main goal.
  • Movement: how often an item is received, moved, consumed, or shipped. Items that move a lot have more chances to be miscounted or mislocated, so ranking by movement aims your counts at the places errors are most likely to happen.
  • Criticality: how badly a shortage would hurt. A low-cost ingredient that stops a production line when it runs out may deserve A treatment even if its value is small.

Many teams start with value or movement, then manually move a few critical items up a class. That hybrid is fine, as long as you write down why each exception exists so the next person understands it.

Whichever basis you choose, use at least a few months of history, and ideally a full year if your business is seasonal. A ranking built on one busy month will mislead you for the other eleven.

How to draw the class lines

Once you have a ranked list, you need to decide where A ends and B begins. There is no single correct answer, but there is a simple process.

  1. Export your SKUs with the measure you chose, such as annual usage value or number of movements.
  2. Sort the list from highest to lowest.
  3. Add a running total and a running percentage of the overall measure.
  4. Mark the point where the running percentage crosses your A cutoff, then do the same for B. Everything below is C.
  5. Look at the result with someone who knows the floor. If a class feels wrong, adjust the cutoffs or move specific items, and note why.

Keep the A class small. If a third of your SKUs end up as A items, the class has lost its meaning, and your counting schedule will show it.

Set a count frequency for each class

The next step is deciding how often each class gets counted. A common starting point looks like this, but treat it as an example, not a standard:

  • A items: once a month.
  • B items: once a quarter.
  • C items: twice a year.

Adjust these to fit your business. A food or supplement producer with short shelf lives and frequent lot changes may want to count A items more often. A warehouse holding slow, stable stock may get away with less. The goal is a frequency high enough to catch errors while they are still traceable, and low enough that the team can actually finish the work.

Then check the math against your labor. Multiply the number of items in each class by its yearly frequency, add the three totals, and divide by the number of working weeks in your year. That gives you the number of item counts per week. If the number is more than your team can handle, lower a frequency or shrink a class before you start, not after the schedule has already slipped.

Build a weekly schedule

Here is a worked example. Say you have 1,000 SKUs, ranked into 100 A items, 250 B items, and 650 C items. These numbers are for illustration only.

  • A items counted monthly: 100 × 12 = 1,200 counts a year.
  • B items counted quarterly: 250 × 4 = 1,000 counts a year.
  • C items counted twice a year: 650 × 2 = 1,300 counts a year.

That is 3,500 item counts a year. Spread across 50 working weeks, it comes to 70 counts a week, or about 14 a day over a five-day week. Now you can test that against reality. How long does one count take in your building, including walking time? Who has room in their shift?

A few habits make the schedule easier to keep:

  • Mix classes in each week, so no single week is all A items.
  • Group counts by area where you can, so counters are not crossing the building for every item. Clear location names help a lot here; our post on building a warehouse location naming system covers the basics.
  • Put counts on a fixed day and time, and give one person ownership of making sure they happen.
  • Leave slack for recounts. Big variances need a second look, and that takes time out of the same week.

In Swift WMS, cycle counting is a Pro feature with a By ABC Class count type, so a class can become a count you schedule and assign to a team member. The ABC Classification page shows which SKUs sit in each class.

Review your classes over time

ABC classes are not set once and forgotten. Products launch, customers change, and seasonal items swing between busy and idle. A class list built last year can quietly send your team to count the wrong things.

Review the ranking on a regular schedule, such as every quarter or every six months. Look for items that have moved a long way up or down the list, new SKUs that have not been classified yet, and items that have gone inactive but still take up count time.

Also look at what your counts are telling you. If a B item keeps showing variances, count it more often, whatever its rank says. Cycle counting is a feedback loop, and the classes should respond to what the counts find. When you post a correction, record a reason. In Swift WMS, adjusting a license plate requires one on every tier, and the Adjustments report on Core and Pro shows where the same problems keep coming back.

Common pitfalls

ABC analysis is simple, which makes it easy to get slightly wrong. Watch for these:

  • Ranking only by value and ignoring critical low-cost items that can stop production.
  • Making the A class too large, so the schedule can never be met.
  • Building classes from too short a window of history, especially in a seasonal business.
  • Never revisiting the classes after the first setup.
  • Counting on schedule but posting variances without looking into them, which hides the root causes the counts were meant to find.
  • Treating C items as items you never need to count. They still need at least one count a year to keep the long tail honest.

The takeaway

ABC analysis turns cycle counting from “count something every week” into a plan. Choose a ranking basis that matches what matters in your operation, keep the A class small, set frequencies your team can actually meet, and check the math before you commit. Then review the classes as your business changes.

If you are still deciding whether to replace an annual count at all, start with our comparison of cycle counting vs. physical inventory. To see how count types, scheduling, and adjustments fit together in one workflow, visit the cycle counting feature page.

How Swift WMS does itCycle counting