Think Build Implement Repeat
London, UK +44 7367 067226
WhatsApp FOLLOW f in X
  1. Home
  2. Blog
  3. Our Demand Forecasting Is Basically One Buyer's Gut Feel. What Happens When She Retires?
Problems We Solve

Our Demand Forecasting Is Basically One Buyer's Gut Feel. What Happens When She Retires?

When one experienced buyer forecasts demand from memory, the business depends on them. We capture their judgement in a forecast model the whole team can use.

Updated 3 min readBy SpiderHunts Technologies

Free estimateNo obligation

Get a free estimate

Tell us what you need. A senior engineer reads every enquiry.

Takes under a minute. We never share your details.

  • Free consultation
  • No commitment
  • NDA on request

Prefer to talk? Book a free 30-minute call →

Quick answer — TL;DR

When demand forecasting lives in one experienced person's head, the risk is not that their judgement is bad but that it cannot be passed on. A machine learning forecast built from your sales history, and shaped by that person's knowledge of promotions, seasons and customers, gives the team a starting number for every product that anyone can review and adjust.

She just knows

Every week your buyer looks at a spreadsheet of stock levels, glances at last year, thinks about which customers are about to place their seasonal orders, remembers that the red version never sells in spring, and places orders. She has done it for years and is usually right. When she is on holiday, someone else tries to follow her approach and the warehouse ends up with the wrong things.

Now she has mentioned retirement, or cutting her hours, and you have realised that nobody else can do what she does. The knowledge is not written down anywhere. There is no formula to hand over. Demand forecasting in your business is one person's memory and instinct, and it is about to walk out of the door.

Why this knowledge never got written down

Experienced buyers work from patterns they have absorbed over years: seasonality, which customers order early, the effect of a price change, the products that sell together. Much of it is genuinely hard to explain. Ask them why they ordered a certain amount and the honest answer is often "it felt right".

Businesses also rarely ask. While the buyer is there and results are fine, there is no pressure to formalise anything. The spreadsheet grows columns only she understands. The process works until the day it has to be done by someone else.

What the dependency costs

While forecasting depends on one personThe effect
Holidays and sicknessOrders placed by someone guessing, leading to gaps or excess
SuccessionNo way to train a replacement except years of shadowing
Growth in product rangeOne person can only give real attention to so many lines
No record of reasoningWhen a forecast is wrong, nobody can learn why
Hard conversationsChallenging the buyer's numbers feels like challenging the person

There is also a quieter cost. Because one person can only focus on so much, the top sellers get careful attention and the long tail of slower lines gets ordered by habit.

How we turn one person's judgement into a shared forecast

  1. We pull your sales and stock history from your ERP, stock system or Shopify, together with price changes, promotions and stockout periods, so the model knows when low sales meant low stock rather than low demand.
  2. We spend time with the buyer and write down the rules she applies: seasonal patterns, customer ordering habits, products that move together, events that matter. Many of these become inputs to the model.
  3. We build a forecasting model per product, or per group of similar products, that predicts demand for the coming weeks with a range rather than a single number, so low-confidence forecasts are visible.
  4. We backtest it against the buyer's own past decisions to show where the model does well and where her judgement still adds something the data does not capture.
  5. We present the forecast in a simple screen or spreadsheet with a suggested order quantity, the reasons behind it, and space for the buyer or her successor to override with a note.
  6. We log overrides and outcomes, so over time the business learns which adjustments help and the model is retrained with that knowledge.

The goal is not to replace the buyer's judgement with a black box. It is to give her a starting point for every line, and to capture enough of what she knows that someone else can pick it up.

What changes for the team

Every product has a forecast, not just the ones the buyer has time for. When she is away, whoever covers starts from the same numbers and the same reasoning. Her adjustments are recorded with a note, which becomes a training record for whoever takes over.

And the conversation changes. Instead of "why did you order that many?", it becomes "the model suggested this, you adjusted it for that reason, and here is what happened". That is a much easier thing to learn from.

For the buyer, it usually means less time on routine lines and more on the decisions that need her experience, such as new products, big customer negotiations and supplier problems. For the business, it means succession is a hand-over rather than a leap of faith.

Is this your situation?

  • One person decides what stock to order, largely from experience.
  • Orders go wrong when that person is away.
  • Nobody could explain the buying logic well enough to train someone else.
  • You have several years of sales history in a system or spreadsheets.
  • A retirement, reduction in hours or resignation is on the horizon.

FAQ

Frequently asked questions

The questions readers ask us after this guide.

Still have a question?

Ask us directly — a senior engineer will get back to you.

Ask about your project

How much sales history do we need?

Enough to see each season at least a couple of times is helpful. With less, the model leans more on similar products and the buyer's rules.

Will the model be better than our buyer?

Not necessarily on every line. It is usually most useful on the many lines she cannot give close attention to, and as a consistent starting point for her successor.

Do we need a data warehouse?

No. We can work from exports of your stock and sales systems to begin with, and automate the data feed once the model proves useful.

What drives the cost?

The number of products, how clean the history is, and whether the forecast needs to feed directly into your ordering system.

What do you need from us?

Sales and stock history, a record of promotions and price changes if you have one, and time with the buyer.

Keep reading

More on Problems We Solve

Start here

Tell us what you are guessing today

Describe the decision, the data you keep and how it is made now. We will tell you honestly whether machine learning would help, what it would need, and if a simpler fix would do the job, we will say so.

  1. You tell us what you needTwo minutes on the form, or a message on WhatsApp.
  2. A senior engineer reviews itAnd comes back with questions, a realistic range and an honest view on fit.
  3. Free 30-minute scoping callWe talk through scope, options and a realistic estimate — with no obligation.
Free estimateNo obligation

Talk to someone who builds this

Send a short brief and we will come back with an honest view and a realistic range.

Takes under a minute. We never share your details.

  • Free consultation
  • No commitment
  • NDA on request

Prefer to talk? Book a free 30-minute call →