Towards a better approach for supplying Jump modems

Every day I receive emails from at least one of our 300+ Jump delivery partners saying they have run out of modems. Our goal is for this never to happen and despite the supply chain challenges caused by Covid, we have tried a variety of approaches to smooth the flow.

We appreciated the support of many of our partners in March this year to redistribute modems they were holding to partners who were desperate for further supplies. But after a month of modem shuffles we became reliant again on the supply of new stock from Ingram Micro.

At any point we have up to 2000 modems in the pipeline, between the time of ordering and the time they arrive at a delivery partner. The challenge is to ensure that supplies arrive, where they are needed and when they are needed, without spending too much time sitting on partners’ shelves.

So today, we have implemented a new approach, called dynamic triggering. Historically our approach has been to set fixed re-supply trigger levels – when a partner’s stock dropped to the trigger level, we automatically placed an order for more supplies. This worked reasonable well when re-supply times were measured in terms of days, but now that we face up to 3 weeks (or at times even longer) to replenish stocks, this approach is not working that well.

Our new approach is to continuously adjust the trigger level, based on historical data. The trigger level for each delivery partner is now based on the number of modems they have issued during the last 3 weeks. For example if a partner has issued 10 modems during the last three weeks, their trigger level becomes 10. Assuming a relatively smooth distribution of modems, the partner should have adequate supplies to last for a further 3 weeks. If delivery times reduce, we can adjust the trigger level to a shorter period, say 2 weeks.

Partners can monitor this themselves in their Jump registers by viewing cell G2 (the dynamic trigger point) compared to their current stock levels (cell H2). When the stock level equals the trigger point, a new order is generated.

We recognise that customer demand will not always be as smooth as the above algorithm requires, but we think this is worth a shot. In particular we expect this will drive supplies to the partners with high turnovers, while minimising the amount of stock held by partners with low turnovers.

BUT, and there is one very big BUT, this does depend on partners ensuring that Profile Forms are completed for every modem issued. We need these to record the number of modems issued, which directly impacts when new orders are placed.

My measure of success will be the drop in emails to jump@diaa.nz requesting further supplies.