01
Forecast accuracy
The most advanced methods, proven against yours.
34.0→
18.6WAPE %
The most accurate forecast is not always the cheapest.
We show you which forecasting approach is worth adopting for your range, measured in your own money, on your own data, before anything changes.
The problem
Forecasting is chosen on accuracy. Inventory is paid for in cash.
Every planning team measures forecast error. Almost none can say what a lower error was worth. Error does not pay for stock sitting in a warehouse, and it does not recover a sale lost to an empty shelf.
So the question that matters never gets asked. Not which approach forecasts best, but which approach leaves you better off once the inventory it holds, the service it delivers and the cost of running it are all counted.
0.Database connections to open, interfaces to build, or changes to your planning systems before you can see the answer.
2.Editions, one product. Restricted runs entirely offline; Standard fits conventional enterprise security.
1.Question answered thoroughly, rather than many approximately. We are explicit about where our answer stops.
Champion vs challenger
One forecast error, three places it shows up
Accuracy is not the point on its own. It is the first link in a chain that ends in cash, and most of the cost sits in the links nobody measures.
02
Inventory held
Stock you can release without losing a sale.
9.4→
6.2Weeks cover
03
Cost to serve
The saving, in money your board understands.
100→
71Cost index
What we do
Three things that make the answer trustworthy
01
Grounded in your business, not a reference model.
We do not hold your range up against textbook assumptions, published benchmarks or how other companies are said to operate. We work from what is actually happening within your supply chain, because that is the only standard against which a change to your business can honestly be judged.
02
Measured in money.
Results arrive in the terms your board already uses: availability to customers, working capital tied up in stock, and the cost of running the approach. You are given the commercial consequence of each option rather than a technical score that somebody then has to translate.
03
Weighted by the strength of the evidence.
Every recommendation arrives with how much weight it will carry, so you know which changes are worth making first and how confident to be in each one. Nothing is presented as certain when the evidence will not support it.
We measure the commercial outcome, not the technical score
We judge every option against your present position
We qualify every recommendation by the strength of its evidence
We answer for each part of your range
We work from the data you already hold
01 / 05
When I realised the platform's ability to accurately diagnose our inventory issues beyond forecast accuracy, it gave me full confidence.
Group Supply Chain DirectorFTSE 100 MULTINATIONAL
Why this matters
Three things a forecast accuracy report will never show you
- A more accurate forecast can leave you worse off overall. We give you the whole answer to ensure that doesn't happen.
- One approach is rarely right across a whole range. Different parts of your business behave differently, and the right choice moves with them.
- A better forecast can be absorbed before it ever reaches your results. Where that happens, more accuracy will not fix it, and we can show you why.
Solutions
One question, answered for every part of your range
The most accurate forecast is not always the cheapest.
Who it is for
Supply chain and inventory leaders carrying a service commitment and a working capital target at the same time, and the finance partners who hold them to both.
Solutions
Which approach each part of your range should actually use.
Proof, not opinion, for every decision behind your forecast.
01
Forecast Method Selection
Know which forecasting approach each part of your range should actually use.
Most organisations run one approach across everything, or several chosen years ago for reasons nobody now remembers. Both leave money on the table, in different places.
We establish which approach performs for each part of your business and what adopting it would be worth, expressed in availability and working capital rather than in error.
- A recommendation for each part of your range, not a single global winner
- The commercial value of each one, in terms your finance function recognises
- A clear sense of how much weight each recommendation carries
02
Measured Against What You Do Today
The only comparison that answers the question you are actually asking.
A benchmark nobody uses proves nothing. Every result we give you answers one question: would this be better than what we do now? That is a higher bar than performing well against a published standard, and it is the only bar that justifies changing anything.
- Every option judged against your present position, not an external reference
- Identical treatment for every option, so the comparison is fair
- A commercial verdict rather than a technical ranking
03
Range Profiling
See which parts of your range are quietly behaving differently.
Few organisations have a clear picture of how differently the parts of their range actually behave, and that difference is usually where the money is. We give you that picture, drawn from your own business rather than from the categories your systems happen to use.
- A view of your range built from how it behaves, not how it is labelled
- The basis on which every later recommendation is made
- A perspective that often surprises the people who know the range best
04
Service and Working Capital Outcomes
What each option would actually have cost you, and delivered.
A forecast only matters through the decisions it causes. We show you the availability and the stock position each option produces, so the choice is made on outcomes rather than on a score, and the trade-off between service and working capital is visible rather than assumed.
- Availability and stock position for every option, on identical terms
- The trade-off between service and working capital made explicit
- Results you can take to a finance partner without translation
05
Decision Confidence
Know how much weight each recommendation carries.
A ranking without a sense of its reliability is an opinion. Everything we give you is qualified by the strength of the evidence behind it, so you can act decisively where the case is strong and proceed carefully where it is not. That is what makes the output safe to take into a boardroom.
- Recommendations qualified by the strength of the evidence
- A clear order of priority for where to act first
- Nothing presented as certain beyond what the evidence supports
06
Reproducibility and Audit
Any result can be reproduced and defended months later.
Every piece of work records what went into it and how it was configured. When a decision is questioned a year on, the answer is a record rather than a recollection.
- A complete record behind every result
- Results that can be reproduced exactly
- An audit trail your governance function can inspect
07
Advanced Diagnostics
For parts of the range where the usual approaches underperform.
Some demand resists forecasting, and knowing that is worth as much as a better forecast. Our deeper diagnostics establish what is realistically achievable for a given part of your range, which stops good money being spent chasing an improvement that was never available.
A recommendation for each part of your range
Judged against your present position
Measured in availability and working capital
Qualified by the strength of the evidence
Reproducible months later
01 / 05
Industries
The same question, asked where it costs the most.
Every sector forecasts, and every sector pays for error in its own currency. We judge forecasting approaches by what they cost in yours, against what you do today.
Builders merchantsFood and drinkFoodservice and grocery wholesaleFMCG and consumer goodsRetail and hospitality rolloutLogistics and 3PLIndustrial manufacturingConstruction
Terraced quarry from directly above
01
Builders merchants and building products distribution
When value holds and volume falls, purchasing plans are running against the wrong number.
This sector publishes its own figures, monthly, and they have been telling a consistent story: sales value holding up while volumes fall. A purchasing plan built on value will over-order. One built on last year's volume will under-serve. Most branches are planning against one of those numbers without knowing which.
We test both, on your own branch and category history, and show which forecasting approach would have left you with the right stock at the right branch, measured in the terms your finance director already uses rather than in forecast error.
What we would test first
- Whether your plans are following value or volume, and what each has cost in stock and in lost sales
- Which approach performs for heavy-side, light-side and seasonal categories, branch by branch
- What is realistically achievable for slow lines before more stock is bought to protect them
Striped fields from the air
02
Food and drink manufacturing
Where forecast error becomes waste within days, and the customer measures your accuracy for you.
Short shelf life turns every over-forecast into write-off and every under-forecast into a missed order. Supermarket supply agreements often add a second cost: an accuracy obligation with a penalty attached. Both are paid in cash, and neither appears in the error report.
Growth makes it worse, not better. A range that doubles in two years carries production rules, run lengths and safety stocks that were set for a smaller business. Revenue can rise while margin goes the other way, and the forecast is usually blamed for a problem that sits in the rules built on it.
What we would test first
- Which forecasting approach each product family should actually use, judged against the one you run today
- What your present production and stock rules cost in waste and in missed service, in your own figures
- Whether the accuracy your customers require is achievable for each line, before more is spent chasing it
The weekly run, from above
03
Foodservice and grocery wholesale
Thousands of lines, three temperatures, and a customer base that changes every week.
Wholesale demand is not one problem. It is a chilled problem, a frozen problem and an ambient problem, each with its own shelf life and its own cost of being wrong, served to customers who open, close and switch supplier without warning. One forecasting approach across the whole book is almost never right for all of it.
The question that matters is not how accurate the depot forecast is. It is how much stock each approach would have had you hold, what it would have cost to hold it, and how many orders it would have fulfilled. We answer that per category and per depot, on your own order history.
What we would test first
- Which parts of the range behave alike, and which only look alike because of how they are coded
- The service and working capital each forecasting approach would have delivered by temperature band and depot
- Where customer churn, not the method, is the real limit on accuracy
Container port at night, aerial
04
FMCG and consumer goods distribution
Wide ranges, export markets, and stock that now has to be in two places at once.
A distributor carrying thousands of SKUs across several markets rarely has one forecasting problem. The fast lines behave differently from the long tail, the home market differently from export, and since Brexit many businesses hold a second stock position on the Continent to serve customers they once served from one warehouse.
Each of those positions is a working capital decision as much as a forecasting one. We show what each approach would have held, where, and what that would have cost against what you do today, so the choice between availability and cash is made deliberately rather than by default.
What we would test first
- Which forecasting approach suits the fast lines, the long tail and the export book, separately
- The cost of dual-territory stock under each approach, in availability and working capital
- How much weight each recommendation carries before any range or warehouse change is made
The estate seen whole
05
Multi-site retail and hospitality in rollout
New sites with no trading history, fresh production, and rotas that have to be set before the first customer arrives.
A site that opened last month has no past to forecast from. Yet it needs a fresh production plan, a stock order and a labour rota on day one, and every one of those is a cost committed before demand is known. Operators usually borrow a forecast from a similar site and hope the similarity holds.
We make that borrowing explicit and test it: which existing sites a new one actually behaves like, how quickly its own history becomes more useful than the proxy, and what each approach would have cost in waste, availability and labour across the estate you already have.
What we would test first
- Which open sites genuinely resemble each new one, by behaviour rather than by format or postcode
- The waste, service and labour cost of each forecasting approach across the existing estate
- When a new site's own history should take over from the proxy, and what it costs to switch too early or too late
Routes as flow
06
Logistics, transport and 3PL
Capacity booked, fleet committed, contracts priced: all on a forecast of someone else's demand.
A logistics business forecasts volume it does not control. Customers' demand decides how many vehicles, shifts and square metres are needed, and the forecast decides what is committed in advance. Too little and service and overtime suffer. Too much and utilisation falls on assets that are paid for regardless.
The right question is which forecasting approach, by contract and by lane, would have produced the best combination of utilisation, service and cost against what you plan with today. That is a commercial answer, and it is the one that decides whether a contract is profitable.
What we would test first
- Which approach performs per contract and per lane, judged against your current planning method
- Utilisation, service and overtime under each approach, on your own volume history
- Where customer behaviour, not the method, sets the ceiling on what can be planned
Engineering at sea scale
07
Industrial and precision manufacturing
Long lead times, finite capacity, and customers whose schedules move faster than yours can.
When material is ordered months ahead and capacity cannot be added in a week, the forecast is not advice. It is a commitment. Customer schedule changes arrive inside the lead time, and the choice is between carrying stock and capacity against them or carrying the risk of missing them.
We show what each forecasting approach would have committed, what it would have cost to hold, and how often it would have been caught out by schedule volatility, so the trade-off between inventory, capacity and service is seen before it is made.
What we would test first
- Which forecasting approach suits make-to-stock, make-to-order and spares, separately
- The inventory and capacity each approach would have committed inside the lead time
- How much of your schedule volatility is forecastable and how much simply has to be buffered
The finished skyline, from above
08
Construction and contracting
The highest-value forecasting problem in the business, in the sector least able to afford getting it wrong.
Project profitability, work in progress and cash are forecast on every job, and the forecast is usually a spreadsheet owned by whoever ran the last one. Margin moves with material prices, programme slippage and variations, and a business can be busy, growing and insolvent in the same quarter.
We apply the same standard here as everywhere else: judged against how you forecast today, measured in money, and qualified by the evidence. Retail fit-out is a particular focus, because its demand comes from the rollout operators we already serve and the two forecasts have to agree.
What we would test first
- How your current job forecasts have performed against outturn, by project type and by stage
- What each forecasting approach would have told you about WIP and cash, and how early
- Where the evidence is strong enough to change how jobs are priced, and where it is not
Across every sector
What does not change from one industry to the next
- Every result is judged against your present position, not a sector benchmark
- Every result is measured in availability, working capital and the cost of running the approach
- Every recommendation arrives with how much weight it will carry
- Evaluation needs extracts of data you already hold, provided as files, and nothing else
Platform
Two editions, one product.
The choice follows your security policy rather than ours.
Deployment options
EDITION 01
Restricted
Runs entirely offline and suits organisations that permit no outbound connectivity at all, including fully isolated environments.
- No outbound connectivity
- Fully isolated environments supported
EDITION 02
Standard
Suits organisations with conventional enterprise security and adds central identity, automatic update checking and audit forwarding into your own monitoring.
- Central identity
- Automatic update checking
- Audit forwarding into your own monitoring
Security and compliance
Designed to support a certified information security management system.
The product is built to be deployed inside an ISO 27001 certified environment without compensating controls. It requires no credentials, runs without elevated privileges, writes only where you direct it, and is delivered as a signed artefact you can scan with your own tooling before it goes anywhere near your estate.
ISO 27001Deployable inside a certified environment without compensating controls
No credentialsRequires none, and runs without elevated privileges
Writes only where you direct itNothing lands on your estate that you did not place
Signed artefactScan it with your own tooling before it goes anywhere
How it fits
Nothing to integrate before you can see the answer.
Evaluation needs extracts of data you already hold, provided as files. There is no database connection to open, no interface to build, and no change to your planning systems required to find out what the answer would be. That is deliberate: the cost of finding out should never be the reason not to.
01
Extracts you already hold
Provided as files. No database connection to open.
02
No interface to build
Nothing to connect, nothing to configure in your planning systems.
03
See the answer
Find out what it would be before anything changes.
Why us
Why not just measure accuracy?
Because accuracy and cost can disagree, and when they do, accuracy is the one that misleads.
WHAT ACCURACY TELLS YOU
A better forecast changes how much stock you carry.
Carrying less saves money, and it also moves you closer to the point where a customer finds an empty shelf.
WHAT IT CANNOT
Whether you finish ahead depends on which effect is larger, and no error metric can tell you.
It has to be measured in money.
This is not a theoretical concern. It is the common case rather than the exception, and it is routinely missed because the report that would reveal it is not the report anybody runs.
What we do differently
01
We measure the commercial outcome, not the technical score
02
We judge every option against your present position, not an external benchmark
03
We qualify every recommendation by the strength of its evidence
04
We answer for each part of your range, because no single approach suits all of it
05
We work from the data you already hold, with nothing to integrate before you see the answer
We answer one question thoroughly rather than many approximately, and we are explicit about where our answer stops.
What we do not do
We are not a planning suite. We do not schedule production, manage suppliers or replace your sales and operations process.