The most common mistake in fixing a warehouse is fixing the wrong thing first. Somebody buys new shelving when the real problem was unscanned returns, or hires another packer when orders were sitting in a queue waiting to be picked. An audit is how you avoid that. It doesn't need to be elaborate. A focused week of looking, counting and timing will tell you more than months of guessing.
To audit a small ecommerce warehouse, spend one day pulling data (orders, inventory adjustments, mis-ships, labor hours), one day following orders through the building, one day running a blind count on random locations, one day timing picking and packing, and one day reviewing receiving and returns. Then score each problem you found by impact and effort, and fix the high-impact, low-effort ones first.
Before You Start
Pick a normal week, not the week before a launch or during peak season. Tell the team what you're doing and why. People work differently when they think they're being judged, and you want to see the operation as it normally runs. Make it clear you're looking at processes, not people.
You'll need: access to your store and inventory reports, a phone for photos and timing, a printed floor sketch (rough is fine), and a notebook or spreadsheet to record everything.
Pull the Data
Start at a desk. Gather the last three months of:
- Orders shipped per day, and the time each order was placed versus the time it shipped.
- Inventory adjustments: every manual change to stock levels, with the reason if one was recorded.
- Mis-ships and damage claims: every reship, refund or complaint caused by fulfillment.
- Labor hours spent on fulfillment, including your own.
- Sales by SKU, including how many times each SKU was picked (order lines, not revenue).
- Returns received and how long each took to process.
From this you can calculate your baseline: orders per labor hour, same-day ship rate, mis-ships per 1,000 orders, and adjustments per week. Write them down. They're what you'll measure your fixes against.
Follow the Orders
Pick five orders of different types (single item, multi-item, fragile, bundle, anything with a special insert) and physically follow each one from the moment it appears in the queue to the moment it's on the truck. Watch, don't help.
On your floor sketch, draw the path the picker walks for each order. Note every time someone:
- Searches for an item instead of going straight to it
- Walks back to the same area twice
- Waits for a printer, a person, a box or a decision
- Handles an item without scanning it
- Asks someone else a question
By the end of the day, your sketch will show you where the walking and waiting happen. It's usually obvious once it's drawn.
Run a Blind Count
Choose 50 locations at random. Have someone count what's physically in each one without seeing what the system says (that's what makes it blind). Then compare.
For every mismatch, try to find out why. Was a return put back unscanned? Were two similar SKUs mixed? Was stock moved and not updated? The causes matter more than the percentage, because they tell you which process to fix.
While you're walking, also note units sitting outside of a labeled location, and shelves full of products that haven't sold in six months.
Time the Work
With a stopwatch, time 20 orders of each common type through picking and packing separately. Record the average and the slowest. Then time the smaller tasks that eat the day: printing labels, building boxes, restocking the packing station, end-of-day manifesting.
Compare picking time to packing time. If picking takes much longer, the issue is layout and walking. If packing takes longer, look at the pack station: are materials within reach, is the printer reliable, does every order need a decision about box size?
Be careful with conclusions from a single day. If one number looks surprising, re-time it on another day before acting on it.
Check Receiving and Returns
These are the parts of the operation that get the least attention, and problems here show up later as inventory errors and stockouts.
- Receiving: how long does it take from a delivery arriving to the stock being sellable in your store? Is inbound counted against what you ordered? Where do shortages and damage get recorded?
- Returns: how many are waiting to be processed, and how old is the oldest? Is there a written rule for what can be restocked?
Ask the team what frustrates them most. The people doing the work every day usually know exactly what's broken. They're often just waiting for someone to ask.
Turn Findings Into a Fix List
By Friday afternoon you'll have a long list of problems. Don't try to fix them all. Score each one from 1 to 5 on two things: how much it would improve speed, accuracy or cost (impact), and how hard it would be to fix (effort). Then sort.
Re-measure your Day 1 baseline a month after making the first round of fixes. If the numbers moved, keep going. If they didn't, the fix probably addressed a symptom rather than the cause. For ideas on what to fix, see our warehouse efficiency guide and the seven warning signs most audits turn up.
Frequently Asked Questions
Spend a week on it: pull three months of order, inventory, error and labor data; follow real orders through the building; run a blind count on random locations; time picking and packing; and review receiving and returns. Then rank what you find by impact and effort.
A blind count is a physical inventory count where the counter doesn't see the expected quantity from the system. It prevents people from unconsciously confirming what they expect to see and gives a truer measure of inventory accuracy.
Orders per labor hour, same-day ship rate, mis-ships per 1,000 orders, inventory location accuracy, inventory adjustments per week, receiving time from delivery to sellable, and the age of unprocessed returns.
A full audit once or twice a year is reasonable for most small operations, ideally a few months before peak season. Track the key numbers weekly in between so problems don't build up unnoticed.
Score each finding on impact and effort. Do high-impact, low-effort fixes first, plan the high-impact, high-effort projects, and skip anything with low impact and high effort. Re-measure after each round of changes.
Not usually for a small ecommerce operation. A structured week of data review, observation, counting and timing will find most issues. Outside help is more useful when the fixes involve major layout changes, new systems, or deciding whether to outsource.
Want a second pair of eyes on your findings?
Run the audit, then send us what you found. We'll tell you which fixes we'd prioritize, and give you a fulfillment quote to compare against the cost of making them.