Your data readiness score
Out of 100, broken down by dimension, with the concrete list of what is missing and what is wrong. It serves you even if you take the next step with someone else.
Phase 0 · Groups and chains of 5+ shops
Four weeks. No cost. If the numbers don't add up, we stop.
You give us a sales history. We measure how good your data is, run a blind backtest over your last 90 days — what BakeWise would have predicted without seeing them — and put a euro figure on the waste and hours you lose. You end with a business case operations and finance can sign. Or a documented “no”.
Read-only access · Nothing to install · No system changes · Data in the EU
What it is and what it is not
What is the four-week diagnostic?
Four weeks on your own sales history, at no cost. We score how good that data actually is against a published rubric. Then we train on everything but the last ninety days, predict that window blind, and put a euro figure on the waste and the coordination hours. You keep all three.
It is
- A measurement on your real data.
- A blind backtest: the model never sees the window it has to predict.
- A business case in euros, with your own figures, for your board.
- A documented “no” if your data does not support it.
It is not
- A demo running on example data.
- A marketing forecast computed over the very period we show you.
- A sales proposal dressed up as a report.
- A funnel you cannot walk out of.
The four weeks
What happens in each week?
One deliverable a week, in order. You get the sales history to us in the format we publish. We score it against the rubric and hand the score back. We train on your history minus ninety days and predict that window blind. We turn the gap into euros.
Week 1: Data delivery
You get the sales history to us — over SFTP, from your ERP, or uploaded by your accountant — in the format we publish. Nothing gets installed on your network.
Week 2: Readiness score
We measure your data against a published rubric and hand back a score out of 100, with what is missing and what is wrong. You keep this even if we go no further.
Week 3: Blind backtest
We train on your history minus the last 90 days and predict that window blind. We compare against what actually happened, product by product and shop by shop.
Week 4: Business case
Avoidable waste, coordination hours and purchasing drift, in euros. A document operations and finance can sign — or reject with data.
This is the product. Not an illustration.
The screens an operations lead uses every morning.
Diagnostic · Phase 0
Readiness score
78/ 100
Pass threshold: 70 / 100
Breakdown by component
- History and continuity24 / 30
- Catalogue discipline (SKU)18 / 25
- Per-shop attribution22 / 25
- VAT coverage14 / 20
Blind backtest · last 90 days
| Shop | MAPE | Avoidable waste |
|---|---|---|
| Pozuelo Estación | 9,4% | €1,180.00 |
| La Moraleja | 11,2% | €940.00 |
| Majadahonda | 13,8% | €760.00 |
| Somosaguas | 10,1% | €880.00 |
The model never sees the window it has to predict. The euro figures are estimates drawn from your own history.
Your data readiness score and the blind backtest, shop by shop.
A reproduction of the real BakeWise interface, with example data.
What you walk away with
What do you keep if we say no?
All three documents, and they are yours either way. The readiness score out of 100, broken down by dimension. The blind backtest result, per product and per shop, with the methodology written down. The business case in euros, with its assumptions in plain sight. They serve you even with another supplier.
The blind backtest result
Accuracy per product and per shop over a window the model never saw. With the methodology written down, so your team can audit it.
The business case in euros
Avoidable waste, coordination hours and purchasing drift, computed on your own history. With the assumptions in plain sight, not in a footnote.
All three deliverables are yours, whether we continue or not.
Entry requirements
Who qualifies?
Five shops or more, with twelve months of digitised sales history at line or receipt level. A named operations lead and a named finance lead, committed to about four hours a week. And a readiness score of at least 70 out of 100 by the end of week two.
At least 5 shops
Below that the diagnostic does not pay off for either side: the right route is Chain, at a public price.
Twelve months of digitised sales history
With receipt or line-level detail. If all you have is daily totals, tell us up front: it changes what we can measure.
A named operations lead and a named finance lead
Committed in writing to around 4 hours a week for the four weeks. Without both signatures the business case goes nowhere.
A readiness score of 70 out of 100 or above
By the end of week 2. Below that we propose a data-cleaning engagement first — which does have a cost — before continuing.
We publish the filter so you can rule yourselves out without spending a meeting.
The readiness score
How is the readiness score computed?
Against five published dimensions, scored out of 100. How granular the data is, whether the months run unbroken, and whether the same product carries the same code in every shop. Then whether every sale knows where it happened, and whether the VAT rates and amounts reconcile.
| Dimension | What we look at | Why it matters |
|---|---|---|
| Granularity | Line-level, receipt-level or daily total? | Without line-level detail there is no per-product forecast, only per-shop. |
| Time coverage | Continuous months, no gaps, no system change mid-way | The model learns seasonality; a system switch breaks it. |
| Product identity | Does the same product carry the same code in every shop? | This is the number-one integration cost. It nearly always has to be resolved by hand. |
| Per-location attribution | Does every sale know which shop it happened in? | Without this there is no per-shop forecast and no per-location P&L. |
| Fiscal and amount consistency | Per-line VAT rates, refunds as negatives, amounts that reconcile | A misattributed VAT rate contaminates the margin and the entire business case. |
How you get the data to us
None of these routes forces a system change or opens up your network.
Available today
- Data delivery by CSV / SFTP in the published format
- Native POS connectors
- Read API with a bk_live_ key
- Delegated upload (your accountant or your vendor uploads the files for you)
- Ingestion of signed TicketBAI / LROE files (Basque Country)
- Native Microsoft Dynamics 365 Business Central connector (cloud)
Native connector within the integration project
- SAP Business One
- SAP S/4HANA
- SAP Business ByDesign
- Microsoft Dynamics NAV
- Sage 200
- Sage X3
- a3ERP
- ICG Manager
- hioPOS
- Odoo
- Holded
- Glop
- Ágora
Product names and trademarks are used for identification purposes only and are the registered trademarks of their respective owners. BakeWise is not affiliated with, sponsored by, certified by, or endorsed by any of these companies. Connector availability and scope are confirmed during discovery.
And if the numbers don't add up
And if the numbers don't add up?
Then we say so, with the data in front of us, and that is where it ends. Not every diagnostic becomes a contract. You keep the readiness score, the backtest and the business case; they serve you either way. We would rather hear a no in week four than run a deployment that fails in month eight.
Frequently asked questions
Is it really free?
The diagnostic is not invoiced. What we ask for is time: around four hours a week from an operations lead and a finance lead across the four weeks. Without that time the result is not good enough to decide on, and then we would both have wasted it.
Exactly what data do you need?
The last twelve months of sales history, with the date, the product, the quantity, the amount and the shop. We publish the exact format and your accountant can read it without calling us. Purchases and stock make the business case more complete, but they are not required to start.
Do we have to install anything or give anyone access?
No. For the diagnostic it is enough to get files to us: over SFTP, exported from your ERP, or uploaded by your accountant. If we later set up a recurring read, it runs on a read-only service account that you create and you revoke.
Who sees our data?
The team running the diagnostic and nobody else. We are a data processor, not a controller: we sign a DPA per customer, host in the EU, and delete your data on request once the diagnostic ends.
What if our readiness score comes out below 70?
We tell you in week 2, with the detail of what fails. It is usually product identity or per-shop attribution, and it is fixable. That clean-up work does have a cost and is quoted separately; we keep it outside the free diagnostic so nobody confuses the two.
We are a chain of fewer than five shops. Can we take part?
The diagnostic would not pay off for you. The right route is Chain, at a public price, with a demo you can watch today without talking to anyone.
Start with the measurement, not with the contract
Four weeks on your real data. If the numbers don't add up, we say so and that is where it ends.