Cuedis

Wholesale: signs a customer is about to stop ordering

In wholesale and distribution, customers seldom leave with a phone call. They do not tell you they found a cheaper supplier or that their own business is struggling. They simply order a little less, a little less often, and one day you realise you have not heard from them in two months.

By then, winning them back is hard. The good news is that the slide is almost always visible in your order history long before it ends. This post explains the signs and how to measure them.

The four early signs

1. Orders getting further apart. A customer who ordered every week now orders every ten days, then every two weeks. Each gap on its own looks like a busy week.

2. Orders getting smaller. Same frequency, smaller amounts. Often the customer is splitting their orders between you and someone else.

3. Fewer product lines. The customer still buys your core items but has stopped buying the add-ons and the higher-margin lines. Usually a competitor is now supplying those.

4. Changes in behaviour. Paying later than they used to, asking for prices on items they never questioned before, more returns or complaints. Not always about leaving, but worth noticing together with the others.

None of these on its own means much. Two or three together, steadily over a couple of months, usually mean the customer is moving business elsewhere.

Why "last order date" is not enough

Many systems can show customers who have not ordered in, say, 60 days. That report is useful, but it shows the ones who have already left. The aim is to catch the ones who are leaving.

It also treats every customer the same way. Sixty days without an order is alarming for a customer who orders weekly, and normal for one who orders four times a year. As with most things, what matters is what is normal for that customer.

Measuring it per customer

For each customer, compare their recent behaviour with their own usual behaviour:

  • Usual gap between orders (for example, the median gap over the past year) against the current gap since their last order.
  • Usual monthly order value against the last two or three months.
  • Usual number of different products per month against recent months.

Here is a starting point for MySQL, comparing each customer's last three months with the nine months before:

SELECT customer, last_order,
       ROUND(recent_per_month, 0) AS recent_per_month,
       ROUND(earlier_per_month, 0) AS earlier_per_month,
       ROUND(100 * recent_per_month / earlier_per_month, 0) AS pct_of_usual
FROM (
    SELECT c.name AS customer,
           SUM(CASE WHEN o.order_date >= CURDATE() - INTERVAL 3 MONTH
                    THEN o.total ELSE 0 END) / 3 AS recent_per_month,
           SUM(CASE WHEN o.order_date < CURDATE() - INTERVAL 3 MONTH
                    THEN o.total ELSE 0 END) / 9 AS earlier_per_month,
           MAX(o.order_date) AS last_order
    FROM orders o
    JOIN customers c ON c.id = o.customer_id
    WHERE o.order_date >= CURDATE() - INTERVAL 12 MONTH
    GROUP BY c.name
) per_customer
WHERE earlier_per_month > 0
ORDER BY pct_of_usual;

On SQL Server, the query is the same except for the dates: replace each CURDATE() - INTERVAL 3 MONTH with DATEADD(month, -3, CAST(GETDATE() AS date)), and CURDATE() - INTERVAL 12 MONTH with DATEADD(month, -12, CAST(GETDATE() AS date)).

The customers at the top of the list are spending the least compared with their own past. Not all of them are leaving, but every one of them is worth a call.

Account for the obvious first

Before calling, rule out the usual explanations:

  • Seasonality. A school canteen orders less in school holidays. A caterer orders more before festivals. Compare with the same months last year where you can.
  • Their own business. A restaurant that renovated for a month, a shop that closed one outlet.
  • Your own stock. If you were out of stock on their usual items, the drop is your problem, not theirs.
  • Changes on your side. A price increase, a new delivery schedule or a new salesperson can all change ordering patterns.

What to do with it

The value is in acting early, while the relationship is still warm:

  1. Call, don't email. "We noticed your orders have changed recently, is everything all right?" is a reasonable question from a supplier who pays attention.
  2. Ask about the lines they stopped buying. That is usually where the competitor got in, and where you can respond: a better price, a delivery change, a missing product.
  3. Check your own service. Late deliveries, short shipments and invoice errors drive customers away quietly. The customer may not mention them unless asked.
  4. Keep a note of what you learned, so the next salesperson knows the history.

Noticing it without a weekly report

The difficulty is not the analysis. It is doing it for every customer, every week, and noticing the slow ones. A customer whose orders shrink 5% a month never looks alarming in any one week.

Cuedis does this as part of its daily reading. It learns each customer's usual ordering rhythm and amounts, watches for slow drifts separately from one-off changes, and mentions a customer in the morning briefing when their orders have been falling against their own normal, with the dates and orders behind it. You can then ask follow-up questions in plain words, such as "what did this customer stop buying?"

Customers who are drifting away rarely tell you. Their orders usually do, months before they stop.