The diagnostic

Find out where the 6.2% goes in your business.

A fixed-fee analysis of your own sales, stock and order data, for retailers. You get a written report of where money is being lost, by cause, with the ten fixes worth most, in rupees, and a backtest of Plavista Fresh against the reorder rule you run today. About six weeks from the day the data lands. Nothing to install, and nothing to buy at the end.

See a sample report  ·  Data specification  ·  A food brand planning launches? See the launch review.

What it looks for

Six places the money usually goes.

Demand your sales report never saw

Lines that sell out early and record the shortfall as low demand, so the forecast learns to order less. We estimate what was actually lost, by line and store.

Stock that exists only in the system

Positive stock on hand, steady expected demand, and no sales for days. Phantom inventory switches reordering off. We list where it is and what it has cost.

Orders placed for the wrong window

Ordering to cover the gap until the next order instead of until the next delivery. It produces the same mid-week gaps, week after week, in the same places.

Suppliers who aren't what the master data says

Stated lead times and fill rates against what actually arrived, and when. The difference shows up as extra stock you are carrying to cover for it.

Waste that comes from ordering, not handling

Expiry and write-offs split by cause: over-ordering, case-pack rounding, slow sell-through. For short-life ranges this is usually the largest single number.

Stock in the wrong place

Excess in one store or DC while another runs out of the same line, and channel imbalances such as quick-commerce orders drawing down store stock.

Inside the report

What you get, shown on a sample chain.

These extracts come from a diagnostic run on a synthetic 25-store chain in Hyderabad (six months of fresh data, May to November 2026). The chain is invented and its problems were planted, so you can see whether the method finds them. Your report has the same sections, on your stores and your numbers.

Sample report · synthetic chain · illustrative, not a client result

1. The ten biggest opportunities, ranked by rupees a month

Each line is traced to a store and a category, compared with stores of the same size, and valued at cost. Waste is valued at what you paid for the stock; lost sales at your margin.

#Store · categoryWhat the data shows₹ a month
1Gachibowli (large) · DairyWaste 8.0% of what was received, against 1.1% at other large stores75,955
2LB Nagar (medium) · DairyWaste 9.6% against 1.7% at other medium stores38,057
3Gachibowli · Fruit & vegWaste 8.4% against 3.8%29,180
4Secunderabad (medium) · all freshSold out on 25.4% of item-days; sales lost at margin ₹28,664 a month against ₹6,445 at other medium stores22,219
5Gachibowli · BakeryWaste 11.3% against 4.4%20,075
6LB Nagar · Fruit & vegWaste 9.9% against 5.6%14,940
7LB Nagar · BakeryWaste 12.8% against 7.3%10,081
8Uppal (medium) · DairyWaste 5.5% against 1.7%9,379
9Uppal · BakeryWaste 16.1% against 7.3%7,313
10Himayatnagar (small) · DairyWaste 7.8% against 3.4% at other small stores7,307
Ten fixes together2,34,506

What it means here: two stores over-order across every category (Gachibowli, LB Nagar), and one under-orders until its shelves empty (Secunderabad). These were the planted problems; the analysis found them without being told.

2. Waste and availability by category

CategoryWaste %Waste a monthSold out %
Fruit & veg6.0%₹4.01 L5.6%
Dairy3.0%₹2.99 L9.1%
Bakery8.7%₹2.64 L6.3%
Chain, 25 stores4.9%₹9.65 L6.6%

Waste % is waste as a share of what was received, at cost; sold out % is the share of item-days that ran out before close. Dairy wastes least but sells out most often: the shelf runs empty before the next delivery. The report shows the same table for every store, against the median store (4.4% waste, sold out on 6.0% of item-days).

3. Stock records that don't add up

In 3 of the 25 stores (Ameerpet, Mehdipatnam, Attapur), the closing stock in the system does not equal opening stock + received − sold − written off on 70–72% of item-days. In the other 22 stores it always does.

Example: Ameerpet, lemons, 28 September. The movements say 111 should be left; the system says 146. Thirty-five lemons the system believes are on the shelf may not be there.

Why it matters: when the record runs high, the reorder thinks there is stock and orders less, and the empty shelf never appears as a stock-out. This is how phantom inventory starts. The fix is a short daily count of the items most likely to be wrong, before any ordering change.

4. Backtest: the same demand, ordered four ways

Who ordersWaste at costDemand served
Store staff, as today₹25.78 L96.5%
ERP min-max reorder rule₹10.98 L95.0%
Plavista Fresh, standard setting₹15.56 L98.1%
Plavista Fresh, set to the rule's availability₹7.03 L95.2%

80 days scored, after settings were chosen on the earlier 100. Illustrative, synthetic backtest (BT-2026-11-13). Published results from live deployments of fresh-ordering software are about 10–18% less waste; your backtest runs on your own history, against the reorder rule your ERP actually uses.

5. What it could be worth: the working

CautiousMiddle
Waste at cost, per store a month₹38,587₹38,587
× share avoided15%25%
Sales lost at margin, per store a month₹6,196₹6,196
× share recovered20%35%
Per store a month₹7,027₹11,816
Chain, a year (25 stores)₹21.1 L₹35.4 L

Every input is shown so your finance team can change it. In your report the starting point is your own audited "inventory written off" line, and the shares come from your backtest, not from these defaults.

6. Limitations and data quality, stated up front

  • Six months of history (16 May to 11 November 2026). It covers one monsoon and one festival season, so there is no year-on-year comparison, and each festival's effect is measured from a single occurrence.
  • Three stores' stock records are unreliable (section 3). Their waste and availability figures are shown with a caution, and the backtest for those stores is less certain.
  • No write-off reasons were recorded, so waste is split into ordering, case-pack and sell-through causes from the order pattern, not from reason codes.
  • No promotion calendar was supplied, so promotion days could not be kept out of normal demand.
  • The audited write-off line was not supplied, so the store-level waste could not be reconciled to the accounts.
  • Where Plavista Fresh did worse, the report says so, store by store and category by category.
How it runs

About six weeks from the day the data lands.

Before we start

Agree scope and fee

A short call to choose the stores and categories. NDA signed. We send a one-page data request your IT team can work from.

Week 1

Data lands and gets checked

We map and check what you sent, and tell you within days what is usable, what is missing, and whether anything needs re-extracting.

Weeks 2–4

Analysis and backtest

Every finding is traced to a product, a store and a date, so your own team can check it against the shelf and the ledger. The backtest replays the months you send, with Plavista Fresh and with your current reorder rule, on the same demand. The method is agreed in writing before your data arrives.

Weeks 5–6

Readout

A written report and a working session with your team: where the money went, by cause, the ten fixes ranked by what each is worth, and how the backtest came out, including anywhere Plavista Fresh did worse.

What we need

Standard exports. No cleaning.

CSV or Excel straight out of your system is fine. The mess is usually where the findings are. Your IT team can work from the one-page data specification.

Twelve months: the full diagnostic

Where the money goes, plus a backtest across a whole year, so every festival, the salary weeks and the monsoon are in it, and seasonality comes from your own history.

Six months: a first diagnostic

Enough to find where the money goes and to backtest those months. Not enough to compare festival seasons year on year: those effects then come from our calendar and are flagged as assumptions in the report.

SalesDaily, by store and product
Stock on handDaily or weekly snapshots, by store and DC
Orders and receiptsWhat was ordered, what arrived, and when
Write-offsExpiry, damage and shrink, with the reason if you record it
MastersProduct hierarchy, pack sizes, shelf life, store list
Questions

What people usually ask first.

Our data is messy. Is that a problem?

No. Every retailer's data is messy, and the patterns in the mess are often findings in their own right: stock records that drift, receipts that never match orders, products with three codes. In the first week we tell you plainly what is usable. If something essential is missing, we say so before the analysis starts, not at the end.

Which systems do you work with?

Anything that can export a file. We regularly see SAP, Oracle, Microsoft Dynamics, Ginesys, GoFrugal, Unicommerce and Tally, and plenty of businesses that run on spreadsheets. For the diagnostic we don't connect to your systems at all; we work from extracts.

How much of our team's time does it take?

A few hours from whoever runs your data extracts, a one-hour kickoff with the business owner, a short check-in during week two, and the readout at the end.

What happens to our data?

It is covered by an NDA signed before anything is shared, stored securely, and seen only by the people doing the analysis. At the end it is deleted or returned, unless you ask us in writing to keep it.

What does it cost, and why isn't it free?

A fixed fee, agreed before we start, based on the number of stores and categories in scope. It is credited in full against your first year if you go on to work with us. We charge because a fee means someone on your side owns the work, and in our experience that is what gets the data extracted.

What happens after the report?

If you want to go further, an optional second stage runs Plavista Fresh side by side in three to five stores for three to four weeks, against matched stores that keep ordering as today, with a stop rule agreed in advance.

Do we have to buy anything afterwards?

No. The report is yours to keep and act on, with your own team, another vendor, or us. Several of the fixes it recommends usually need no software at all.

Tell us a little about your business and we'll come back within one working day.

Ask for a diagnostic