Order from a six-week supplier knowing what your customers will use next month
For industrial suppliers and distributors of fasteners, abrasives, PPE and maintenance consumables. Tens of thousands of items, customers who use the same quantity every month and suppliers who take weeks to deliver. We turn that repeat consumption into a forecast and propose each supplier’s order with its lead time and minimum order.
Book a 30-minute demoHow it works
Read-only connection · Tested on your own data first · Data hosted in the EU · Demos in English or Spanish
The short answer
Inventory management for an industrial supplier turns the repeat consumption of your customers (gloves, cutting discs, fasteners, PPE) into a forecast and proposes each supplier’s order with its lead time and minimum order. It sorts tens of thousands of items into those that rotate, those that sell now and then and those that are dead stock, splits PPE across sizes by what really sells, and explains why each part of the dead stock came in.
By Manuel Gómez González, founder of Arume · Reviewed on
What it does
Repeat customers, forecast one by one
The machine shop, the maintenance contractor and the factory on the industrial estate buy almost the same each month: cutting and grinding discs, carbide drills, gloves, glasses, earplugs, maintenance sprays. We detect which items each customer buys at a steady rhythm and forecast that separately from the more erratic counter sales.
Orders by supplier, with their lead time and minimums
The local wholesaler delivers in 24 to 48 hours, the European tool maker in two or three weeks, imported fasteners in six to eight, by container. Each has its own minimum order and free-carriage threshold. We calculate the forecast demand until the next order can arrive, using the real lead time measured from your past orders.
A catalogue sorted into three groups
Items that rotate regularly get a forecast and an order. Items that sell now and then get a risk calculation. Items that have not sold in a year go on a dead-stock list in money. What you review each week is a short list, sorted by what you lose if you do not order.
Units that match reality
The same screw is bought in boxes of 200, sold by the box to the factory and loose at the counter. Cable is bought on 100 m reels and sold by the metre. We load each item’s conversion table and count everything in the smallest unit, so that stock matches the shelf.
PPE sizes
A safety boot comes in ten sizes and they do not sell evenly. We treat each size as its own item and learn its split from your sales, so the order for a model is shared across sizes by what sells, and sizes that have not moved in months are flagged.
Dead stock with its origin
We list dead stock in money and by age, crossed with why it came in when that can be reconstructed: the customer it was bought for, the supplier, the date. Large one-off purchases for a project are flagged so they do not raise the minimums again.
How are tens of thousands of items ordered today?
A mid-sized industrial supplier has 15,000 to 50,000 items set up: fasteners in every size and grade, hand and power tools, abrasives, adhesives, PPE in every size, welding consumables, maintenance chemicals. Supplier price lists add many more. The person deciding what to order is usually one, two at most, and also answers the counter and the phone.
At that volume nobody reviews item by item. You order what you see is missing, what a customer has complained about and what the supplier’s rep suggests when passing by. The minimums in the item record were set when it was created and almost never touched. The result is well known in the trade: stockouts on what sells most and shelves full of what does not.
Inventory control with AI does not replace that person. Each week it puts the short list of what really needs ordering in front of them, with the quantity calculated and the reason, and the list of what is stuck.
Why are repeat customers the easiest demand to forecast?
Unlike a shop, a good part of an industrial supplier’s sales come from customers who repeat. That repeated consumption is the easiest demand there is to forecast, and almost no warehouse uses it. We detect it customer by customer: which items a customer buys at a steady pace, how often and in what quantity. It is forecast on its own and added to counter sales, which are more irregular.
There is a side effect. When a regular customer stops buying an item they bought every month, the system notices. They may have switched supplier for that product, and your sales rep will want to know before it happens with the rest.
How do you handle suppliers with different lead times?
With lead times this different, a fixed minimum in the item record does not work: it covers the item that arrives in two days and falls short for the one that takes two months. For each item we calculate the demand forecast until the next order of its supplier can arrive, and group by supplier.
- What to order from each supplier now so as not to stock out before it arrives.
- If you are below the minimum order or the free-carriage threshold, which items to bring forward and what it costs to hold them earlier.
- Which items have a faster alternative supplier in case the main one is late.
- A warning on imported items weeks ahead, not when there is no remedy left.
Which rules make PPE sizes matter?
Your customer cannot settle for whatever size is in stock. Council Directive 89/656/EEC on the use of personal protective equipment at work says PPE must be appropriate for the risks involved, take account of the worker’s state of health and ergonomic requirements, and fit the wearer correctly after any necessary adjustment. It also says the employer provides it free of charge and keeps it in good working order through maintenance, repair and replacement.
For a supplier, that turns a missing size into a customer problem. We treat each size as its own item and learn its split from your sales, which depend on your customers: a workshop crew does not have the same sizes as a cleaning team. When a large customer approves a new boot or glove model for its whole workforce, we move the history of the old model to the new one and stop reordering the old when what is left runs out.
Which programs do industrial suppliers use?
Almost every management program for hardware and industrial supply exports sales, stock and supplier orders to Excel or CSV, and that is enough to start. If you use Sage or Odoo, we connect directly with a read-only user. In the first stage we look at what can be taken out of yours and choose the shortest route.
What data do you need, and what does it not do?
Sales per item, customer and date (delivery notes and counter tickets); today’s stock; supplier orders with order and receipt dates; and unit conversions if you have them. With a year of sales we can classify the whole catalogue and forecast the consumption of regular customers. Customer contact details are not needed: a code is enough.
It does not negotiate prices, choose suppliers or decide which customer to keep. It gives you the forecast and the order proposal, and a person approves.
How do you test it and how do you start?
First we clean the data and match the units. Then the system replays your last months without knowing the future, to compare with the orders you placed. After that, one or two months in shadow mode. If you would rather see something before talking, send us a sales and stock export and we return a short report: how many items do not move, how much money they add up to, and which regular customers are buying less than a year ago. The first step is a 30-minute video demo, in English or Spanish.
Frequently asked questions
With 40,000 items, is it too much?
We do not treat all 40,000 the same. First we classify the catalogue: items that rotate regularly get a forecast and an order; those that sell occasionally get a risk calculation; those that have not sold in a year go on a dead-stock list in money. What you review each week is a short list.
Does it take supplier minimums and carriage thresholds into account?
Yes. Each supplier has its real lead time, measured from your past orders, its minimum order and its free-carriage threshold. If you do not reach the minimum, we show what to bring forward from what you will need soon, and what holding it earlier costs.
How do you treat a big one-off purchase for a project?
It is detected as one-off because of its size compared with that customer’s normal consumption of the item, and marked separately. It does not raise the minimum or repeat in the forecast. If the same customer does it every season, then it is learned as seasonality.
Does it work for PPE with sizes?
Yes. Each size is an item and its split is learned from your sales. In the order, the total for the model is divided across sizes by what really sells, and sizes that have not sold for months are flagged so they are not reordered out of habit.
Do we need to change our management program?
No. We read from it with a user that can only read, or from an export. The approved order goes back as a purchase order or as a spreadsheet for the buyer.
Where is the data stored, and what if we leave?
In the EU, with daily backups. If you stop working with us you can export everything, and the code and data stay in your company’s name.