Talk to an expert

Tell us what you need and we will call you back.

🇮🇳 +91

    Stock Audit Explainer

    Why Dark Store Stock Drifts in 48 Hours: Picks, Returns, Damages

    Where the gap between a dark store's system and its shelves comes from: wrong picks, returns and damaged or expired stock, what the pattern of variances says about the cause, and what slows the drift. How a dark store is counted is covered in other guides.

    Quick answer

    Dark store stock drifts within days because wrong picks, unposted returns and unrecorded damage repeat many times a day across thousands of fast-moving lines.

    Causes One wrong pick makes two wrong balances; returns are shelved before posting; damage leaves the shelf before the record.
    Reading it Paired variances mean an accuracy problem; one-sided shortages mean goods have left the site.

    Fix pick errors first, post returns in one step, log damage at the pick face, and measure against an independent baseline.

    In this guide: 6 sections Start reading
    Why dark store stock drifts: Your dark store count keeps drifting.

    Why the Numbers Move So Fast

    Dark store stock records go wrong within days because three ordinary errors, wrong picks, unposted returns and unrecorded damage, happen at a volume few other stock formats see. A micro-fulfilment centre packs thousands of fast-moving lines into a small space and handles each of them many times a day, so a small error rate turns into a visible gap between the system and the shelf very quickly. The 48 hours in the title is not a measured constant. It is shorthand for how soon the drift shows: in days, not at the next quarterly count.

    Out the door

    Picking errors

    The wrong unit, or the wrong quantity, goes into a bag. One mistake leaves two lines wrong.

    Back through the door

    Returns

    Failed deliveries and refusals come back into a site designed to send goods out, and are shelved before they are posted.

    Never through the door

    Damage and expiry

    Crushed packs and short-life goods leave saleable stock on the floor long before they leave the record.

    None of the three is fraud, and none of them is rare. What makes the format different is the rate at which they repeat. A weekly count in a dark store is measuring a stock that has already turned over since the count began.

    Picking Errors at Speed

    The largest single source of drift is the wrong unit leaving the building, and it happens because the site is built for speed rather than care. Two forms account for most of it.

    • The look-alike neighbour. Two products in similar packaging sit side by side. A picker working against a delivery promise reaches for the nearer one, scans the slot rather than the item, and the order leaves looking correct. Nothing in the transaction flags it, because the system was told the right item went.
      Wrong item
    • The miscount under pressure. Two taken where three were ordered, or three where two were. It is far more common with small items picked by hand from an open bin than with anything scanned one by one.
      Wrong quantity

    How one reach becomes two errors: the order asks for item A, the picker takes item B from the next slot. The system deducts A, so A now shows one fewer than the shelf really holds. The shelf loses B, so B now shows one more than the shelf really holds. One mistake, two wrong balances, in opposite directions.

    The second customer to order item B is the one who finds out. The system says there is one left; the picker finds an empty slot, and the order is short or substituted. That is why drift shows up first as pick accuracy and fill-rate problems, well before anyone counts.

    Returns and Failed Deliveries

    Everything in a dark store is arranged to move goods out. Anything that comes back moves against that flow, and three things tend to go wrong with it.

    1. Shelved but not posted. A rider brings back a refused order, and someone helpfully puts the items back where they belong without booking the return. The shelf rises, the system does not, and the next count shows an excess nobody can explain.
    2. Shelved without a condition check. A chilled item that has been out of temperature, or a pack damaged on the round trip, is not the same stock that left. Put back into pick-face stock unchecked, it goes to the next customer and may come back again.
    3. Not shelved at all. Returns pile up near the receiving point, neither in saleable stock nor written off, until someone puts them away in bulk without knowing where they came from.

    Where returns actually end up is worth checking on the floor rather than in the process document. The written process usually says returns are inspected and posted on arrival; the receiving corner often says otherwise.

    Damage, Expiry and Unsaleables

    The third source of drift is stock that is physically present and commercially finished. It builds up faster here than in most formats because so much of the range has a short life.

    • Short-life ranges. Fruit and vegetables, dairy, bakery and chilled lines lose saleability within days. Write-offs are part of normal trading, not an exception, and every one not recorded as it happens leaves the system carrying stock the shelf does not have. Stock rotation by FEFO limits the loss but does not record it.
    • Damage found mid-pick. Logging a crushed pack properly costs a delivery window, so the picker takes the next unit and puts the damaged one aside. The record is updated later, if at all.
    • The write-off nobody approved. Damaged, expired and returned goods are eventually thrown out by site staff, in quantities no one authorised and with no document. The loss is real, the approval is missing, and the gap appears at the next count as shrinkage with no explanation.

    What the Variance Pattern Reveals

    The shape of the variances says more than their total. Reading the shape before deciding what went wrong avoids a great deal of misplaced suspicion, because the two patterns below look identical in a summary and completely different line by line.

    Reading a dark store count line by line

    The same net shortage can mean an accuracy problem or a loss problem.

    Paired variances: an accuracy problem

    • One line short by a quantity, another over by the same quantity
    • Usually neighbours on the shelf, or products that look alike
    • Nothing has left the building: the stock is there, the record is wrong
    • The fix is slotting, labelling or scanning, not security

    One-sided variances: a loss problem

    • A shortage with no matching excess anywhere in the count
    • The goods are not in the site
    • The cause is an unrecorded dispatch, unrecorded damage or disposal, or loss
    • Each needs its own follow-up, and some need an investigation

    A site whose variances are mostly paired has an accuracy problem. A site whose variances are mostly one-sided has a loss problem. The remedies have almost nothing in common, which is why a single network shrinkage figure is a poor guide to what to do next. When one-sided shortages keep returning at the same site, the guide to forensic stock audits explains how auditors test whether stock is being lost deliberately.

    Slowing the Drift

    Drift cannot be stopped in a format this fast, but it can be slowed, and the order of work matters.

    1. Pick errors first. They create two record errors from one mistake, and they respond to physical change better than to instruction. Separate look-alike products, move fast lines to slots that do not need a reach past a neighbour, and give the most-confused lines their own clearly marked place. The paired variances from your counts name the pairs being confused; that list is the work plan.
    2. Returns as one action. Record and shelve a return in the same step, after a condition check, instead of logging it for later. In a site this busy, anything left to the end of the shift tends not to happen.
    3. Damage logged at the pick face. Give pickers a way to mark a damaged unit in a few seconds, and route it to one damage location that is counted and cleared with an approval, so that disposals leave a record.

    Measure the effect against a baseline you trust. Improvement measured against a system figure that was already wrong proves very little, so an independent count before the changes is worth having, across the stores that matter. A stock audit for dark stores establishes that baseline store by store, separating paired errors from genuine loss.

    Questions people ask

    Dark stores: what readers ask

    Short answers to the questions people ask about dark stores. The causes of drift are in the sections above.

    How does a dark store work?

    A dark store is a small fulfilment site closed to walk-in shoppers. Orders placed in an app are routed to the nearest store, picked from the shelves by staff, packed and handed to a delivery rider, usually within minutes. The store is replenished from a larger warehouse, often more than once a day, and returns come back through the same riders. That constant flow in and out is why its stock records drift so quickly.

    Is a dark store a warehouse?

    In function, yes: it stores stock and fulfils orders, and it is closed to customers. It differs from a conventional warehouse in size, location and speed. A dark store sits close to the people it serves, holds a narrower range of fast-moving lines, and picks single orders in minutes rather than shipping pallets or cartons. Many businesses call it a micro-fulfilment centre.

    What is a dark store manager?

    The person who runs one dark store day to day: staffing the shifts, meeting delivery times, receiving replenishment and keeping the stock accurate. On stock, the role usually covers how returns are posted, how damaged and expired goods are recorded and approved for disposal, and how variances found at a count are investigated, which are the three sources of drift described above.

    Are dark stores profitable?

    There is no single answer: it depends on order volume per store, basket size, delivery costs and rent, and it varies by business and city. Stock accuracy is one of the levers. Unrecorded damage, expired goods and short or substituted orders all reduce margin, so a store whose records drift is losing money in ways its own reports do not show.

    Before you fix the drift

    Dark store figures drifted? Start from an independent baseline

    An independent count across the stores that matter, separating picking errors from genuine loss, gives the changes something real to be measured against.

    • Counts planned around the stores' operating hours
    • Variances traced to picks, returns, damage or loss
    • Reports reviewed and signed off by a Chartered Accountant
    Subham Jhunjhunwala

    About the author

    Subham Jhunjhunwala

    CA Finalist at Patron Accounting. Author of this guide and reviewer of the firm's stock audit content. See his profile and other guides.

    Reviewed by
    CA Sundram Gupta, Chartered Accountant
    Regulatory statements
    Practice guidance; no statute is cited on this page
    Published
    19 August 2026

    Talk to a CA

    Tell us about the dark stores and the right person calls you back.

    🇮🇳 +91

      By sending this form you agree to our privacy policy.