How to Scale a Distribution Business

Scale multiplies the exception you have been heroically fixing at 6 p.m. If Monday's wave still depends on one person who knows where the overflow lives, the new territory will miss A items in week two. Tripwires before expansion: 97% fill on A items for eight weeks, 98.5% pick accuracy, dock-to-stock under 24 hours, and a cycle-count program that is not 'we count when we are slow.'

Sales forecasts are a wish. Expansion money should be sized off current turns, current DSO, and the incremental inventory the new lane needs at the current fill promise. If you cannot fund 60–70 days of that without starving the home book, you are not ready. You are fundraising.

The SBA page on managing growth is blunt about cash. In wholesale that page should be taped next to the territory map.

CCC = DIO + DSO − DPO. Worked scale test. Home book: DIO 48, DSO 34, DPO 26 → CCC 56 days. You want a second county that will do $180,000 of monthly sales at a 55% inventory-to-sales ratio on the extra SKUs. Incremental inventory ~$99,000, incremental receivables ~$204,000 at 34 DSO if they ramp in 60 days, payables maybe $70,000. Peak cash out before the new lane pays you is north of $180,000 plus the extra driver's loaded cost. If the line of credit is $150,000 and already 60% used, the map is a fantasy.

Stress the model: sales ramp 30% slower, collections 10 days longer, and one container late. If that stress empties the account, cut the SKU list for the new lane or delay the hire. Optimism is not a covenant.

The Census AWTS inventory-to-sales ratios are a useful outside check when a supplier is pushing you to 'stock the full line' in the new county. Full line is how DIO goes from 48 to 70.

If orders die in purchasing, do not hire a hunter. If the hunter has 40 open doors and no time to expand SKUs, hire a farmer or give inside a penetration list. If the dock cannot ship the orders you already have, overtime and a second picker will return more GP$ than a new logo.

Write seats with one primary number. Buyer: fill and excess. Farmer: SKU-per-door and GP$/stop. Hunter: surviving 90-day doors. Warehouse lead: pick accuracy and cutoff hit rate. If two people own the same number, nobody does.

Owner salary is part of the scale model. A house that only works because the founder is the backup buyer, the backup driver, and the credit manager is not ready for a second building. Document the ugly exceptions — damaged pallet, short ASN, customer who will not accept a substitution — before you clone the operation.

New territory plus 'we should carry what they ask for' is how you add 400 SKUs that turn twice a year. Freeze the open-to-buy for new items during the first 90 days of a lane. If a door needs an item you do not stock, special-order it once, measure reorder, then add. The first miss on an A item you already sold at home will cost more than the lost long-tail order.

Private label and extra brands feel like scale. They are warehouse complexity. Add a brand when the first one is above 97% fill and the buyer can tell you which doors will drop a competitor for it. Otherwise you bought a second set of safety stock.

Satellite warehouses fail when item masters, lot rules, and cutoffs diverge. Clone the item file, the slotting logic for A items, and the same daily exception meeting. If the satellite cannot see home inventory, you will transfer in panic and pay the mileage twice.

3PL overflow is a scale tool, not a personality. Use it for seasonal peaks or a distant cluster you cannot density yet. Bring it back in-house only when the pick profile is stable and the 3PL's accessorials (storage minimums, peak surcharges, order minimums) exceed a small lease plus two people.

Pilot one adjacent county, 40 named doors, 180 SKUs that already turn at home, one inside person plus one shared driver day. Cap inventory transfer at $90,000. Kill criteria at day 75: fewer than 12 doors ordering weekly, fill under 94%, or CCC on the lane above 75 days. If you hit kill, you lost a controlled amount. If you skip the cap and 'fully stock' the lane, you will be discounting the leftovers in November.

Tax and nexus travel with you. A second state is an IRS and state registration event, plus sales-tax permits and, if you run trucks, IFTA and fuel credentials. Scale is not only pallets.

Overtime is a scale signal. If the warehouse is over 12% OT for six weeks, you do not have a culture problem — you have a volume or a slotting problem. Adding a hunter on top of that OT is how you fund both a new payroll and more Saturday. Fix minutes per line first.

Lead times on hiring in wholesale are ugly. A competent buyer or a CDL driver is not a two-week post. Start the search when the tripwire is in sight, not after you announce the county. Temp labor on the dock is fine for a peak; temp labor in purchasing is how you duplicate POs.

Write backup seats. If the founder is still the only person who can release a credit hold or approve a return over $250, you cannot take a week off and you cannot add a building. Push the authority down with a dollar limit and an audit. Scale that still needs the owner's thumb on every exception is a taller stool.

You need one item master, one customer price file, and a credit hold that the dock cannot ship around. Spreadsheets that 'we will migrate later' become two companies that share a logo. If EDI is how the new banner will order, complete a test 850/856/810 cycle on a dummy door before you promise a go-live.

Cycle counts and lot/expiry rules travel. A satellite that does not honor FEFO will ship the new dates and leave you aging product at home. Clone the rule, then audit it in week two with a count, not a meeting.

Insurance and registrations scale with miles. Cargo limits, hired/non-owned auto, and workers-comp class codes change when you add a driver or a state. Call the broker before the lease, not after the first claim. The SBA insurance overview is a start; your category may need product-liability limits a banner will specify in the packet.

Freeze new suppliers during the first 90 days of a lane unless the item is an A-item stockout with no substitute. Freeze custom labels, customer-specific packs, and new EDI maps that are not required to invoice. Every one of those is a second operating system. You can add them when the lane's CCC and fill are boring.

Freeze founder heroics. If the owner is still the backup driver on Thursday, the new county will inherit that expectation. Put a paid backup in the model — even a contract driver one day a week — before you print the map. Scale plans that assume the owner has spare Thursdays are fiction.

Write a one-page 'not yet' list and tape it in purchasing and sales. Not yet: second private label, third building, a marketplace storefront, a 400-SKU expansion pack from a hopeful factory. The list is not permanent. It is how you finish the thing you already started. Houses that scale on five simultaneous bets usually finish none of them with a fill rate anyone would put on a line card.

Wait if A-item fill has been under 96% for a month, if DSO has crept past 40, if the line of credit is over 70% used, or if the backup buyer is still 'the owner on Sunday.' None of those problems get better because a county looks open on a map. They get worse when you add doors.

Wait if the last three new logos are already dual-sourcing you. That is a service signal. A hunter will call it price. Check the misses first. If you are late or short, more logos will dual-source you faster.

Tell the supplier who is pushing a territory expansion the same story with numbers. Brands love a wider map. They do not love a distributor who shorts the existing authorized doors to decorate a new one. A 90-day hold with a fill recovery plan is a better partner conversation than a hopeful launch and a January apology.