Hub-and-spoke vs cross-docking vs hybrid: how to decide for a 50-route operation

The clearest sign a distribution network has outgrown its model is when the operator can name, off the top of their head, the SKUs that never sit still. "The fizzy water and the white bread — they come in at six and go out by ten. They've never seen a rack." And yet they are racked. They get put away. They get picked. They get moved twice, sometimes three times, before they leave the building. Every one of those touches has a cost, and none of them adds value to the customer.

I've watched operations spend a year arguing about WMS upgrades and pick paths when the real answer was structural: stop storing the fast-movers. The shape of the network was wrong for the shape of the demand. Pick paths only help when the goods should have been picked. For the top 30 SKUs in many regional ops, the goods shouldn't have been picked at all — they should have been cross-docked.

This post is about how to decide. Not the textbook three-paragraph "here's hub-and-spoke, here's cross-docking" treatment you can get anywhere. Real thresholds, real trigger conditions, and a worked example for a 50-route regional operation.

The three models, briefly

Hub-and-spoke is the default. Goods come into a central warehouse, get put away into bulk or pick locations, and sit there until orders are placed. When orders come in, they're picked, consolidated by route, loaded, and delivered. This is what most 3PLs and regional distributors do. It absorbs supplier variability, supports any SKU velocity, and gives you a single inventory pool to draw from. The cost is touches and storage square-metres.

Cross-docking strips the storage layer out. Goods arrive already sorted by destination — either by the supplier or by a quick sort on the dock — and ship out the same day or the next day. There's no putaway, no pick, no replenishment. The dock becomes a sort-and-stage area, not a storage building. Done well, the labour-per-line drops by half and the floor empties out. Done badly, you've built a single point of failure that one late inbound can knock over.

Hybrid is what most networks evolve into whether they planned to or not. Storage for the long tail, cross-dock for the fast-movers, with the two flows sharing a building and a team. The split is rarely 50/50 — it's usually 70% of volume cross-docked through 20% of the SKUs, with the long tail handled traditionally. The hard part of hybrid is operational discipline: keeping the cross-dock lane free, not letting fast-movers drift back into racking because someone "had a spare slot".

Sizing thresholds: when each model fits

These aren't laws, but they're the patterns I see again and again across regional distribution ops in the UK.

Under 20 routes per day → hub-and-spoke, almost always. At this scale you don't have the inbound predictability or the SKU concentration to make cross-docking work. Suppliers don't prioritise you. Your top 10 SKUs probably aren't doing 80% of moves yet — your customer base is too narrow. A simple hub absorbs the variance. Trying to cross-dock here usually means one supplier no-show writes off a whole shift.

20 to 80 routes per day → start cross-docking the fast-movers. This is the band where hybrid starts paying. You've got enough volume on a handful of SKUs that bypassing storage on those specifically saves real money. You also have enough leverage with suppliers to ask for pre-sorted, time-windowed inbounds on the lines that matter. You're not cross-docking everything — you're cross-docking the 20% that move every day, while the long tail stays in racking.

80+ routes per day → hybrid or majority cross-dock. At this scale storage cost is brutal and labour throughput is the constraint. Most operations of this size end up with two distinct flows under one roof: a high-velocity cross-dock lane that processes the day's fast-movers in a four-hour window, and a smaller hub-and-spoke operation handling the slower SKUs and customer-specific stock. The buildings start to look different too — more dock doors, less racking, longer marshalling areas.

If you're below 20 routes and someone is selling you a cross-dock transformation, ask hard questions. If you're above 80 routes and still putting every pallet into bulk, you're paying twice for moves you don't need to make.

Six questions to answer before you pick

The thresholds get you in the right neighbourhood. These questions get you to the right address.

1. SKU velocity distribution. Take the last quarter of orders. Sort SKUs by line count, descending. What share of total moves comes from the top 20%? If it's 80% or higher, you have the kind of skewed distribution that rewards cross-docking the top end. If it's flatter — say the top 20% does only 40% of moves — there's no clear cross-dock candidate set and a hub still makes sense.

2. Inbound predictability. Can your suppliers hit a two-hour delivery window, consistently? Not "in theory" — in practice, over the last 30 inbounds. Cross-docking depends on goods being on the dock when the outbound wave needs them. If your top suppliers routinely arrive a half-day late, you can't cross-dock from them no matter how nicely the SKU velocity looks.

3. Customer geography. Are deliveries clustered into a handful of postcodes, or spread thinly across a region? Cross-docking is easier when customer destinations repeat — you can pre-sort by route on the inbound. Spread geography means more bays to stage into, more sort complexity, and a higher chance the cross-dock floor turns into a temporary store.

4. Volume per drop. Full pallets, mixed pallets, or case-pick? Full-pallet drops cross-dock beautifully — the supplier can label by destination and the pallet moves once. Case-pick is the hardest case for cross-docking because you're effectively building a pick from inbound, which adds the labour you were trying to remove. Mixed-pallet drops sit in the middle: workable if the mix is predictable.

5. Lead-time tolerance. Same-day, next-day, or 48-hour? Hub-and-spoke handles all three. Cross-docking really shines at next-day because it gives you a full overnight window to receive, sort, and dispatch. Same-day cross-docking exists but requires near-perfect supplier timing. If your customers tolerate 48 hours, the case for cross-docking weakens because you have time to make the hub work.

6. Storage cost. What's your warehouse costing you per square metre per year, fully loaded — rent, rates, lighting, heating, fork trucks, racking depreciation? In a cheap regional shed, storage is almost free and the cost of touches dominates. In an expensive urban DC, every square metre saved by emptying the racking pays back fast. Cross-docking is much easier to justify when the floor is expensive.

If you answer all six honestly and four of them point the same way, you have your answer.

Triggers to switch from hub to hybrid

Most networks don't redesign overnight. They drift. The question is whether you notice the drift before it costs you a peak season.

The clearest trigger: the top 50 SKUs are now doing more than 60% of moves. When the curve gets that skewed, you have a small set of high-touch lines that would benefit enormously from bypassing storage entirely. Below 60% the case is weaker — the cross-dock lane doesn't get enough volume to justify the operational complexity. Above 60%, you're leaving money on the floor every shift.

A close second: your storage cost has overtaken your transport cost on the P&L. Distribution operations historically spend more on the vans than the building. When that ratio flips, it usually means inventory has piled up — slow-moving SKUs that should have been deflicted or returned, fast-movers that are sitting in bulk longer than they should. Cross-docking the top end is a way to physically refuse to store goods that don't need storing.

Third: drops have shifted from full-pallet to case-pick. Many regional operations started life as pallet-in, pallet-out. Then they took on smaller retail customers, or their existing customers started ordering more frequently in smaller quantities. The pick complexity went up and the warehouse never re-shaped. Hybrid forces the question: which lines still go pallet-in, pallet-out, and which need pick-friendly stocking? The fast-movers in case-pick volumes are the strongest cross-dock candidates because each one is being touched many times a day in the current model.

If two of those three triggers are firing, plan the hybrid move. If all three are firing, you're already behind.

The cross-docking trap

I've seen more cross-dock projects go sideways than succeed on the first try, so it's worth being blunt about the failure modes.

It depends entirely on supplier predictability. Cross-docking is a tight choreography: goods arrive in window, get sorted, get loaded, leave. There is no slack. A storage layer absorbs variance — a missed inbound just means you pick from stock. Without that buffer, every late lorry from a supplier becomes a problem on your outbound. If your suppliers aren't already hitting their windows 90%+ of the time, cross-docking will expose every weakness in your inbound process before it saves you any money.

Labour cost spikes, because there's no slack in the timeline. In a hub, you can flex picking across the day — start early, finish late, adjust to volume. In a cross-dock, the work has to happen in a specific window between the last inbound and the first outbound. That window is often three to five hours. You need the people, you need them in that window, and you need them able to handle the peak. Overtime, agency cover, and weekend rates start to bite. The labour-per-line drops, but the labour-per-hour goes up.

One late inbound can shut down a whole outbound wave. This is the one that catches operators out. In hub-and-spoke, a late supplier delivery is a problem you can manage — you pick around it. In a cross-dock, that delivery might be the only source of stock for an entire region's outbound. If it doesn't arrive by 11am, eight routes don't leave by 1pm. You've turned a supplier problem into a customer problem with no buffer in between.

The fix isn't to avoid cross-docking. It's to be honest about which SKUs and which suppliers can really handle it. The mistake is treating cross-docking as a building-wide strategy. Treat it as a SKU-level decision and most of these problems disappear — because the fast-movers from your most reliable suppliers can cross-dock, and everything else stays in the hub.

A worked 50-route example

Imagine a regional distribution operation running 50 routes per day across the Midlands. 200 active SKUs. Three classes of customer: 40 retail accounts, 80 hospitality accounts, and a long tail of independents. Average drop size of 1.4 pallets. Next-day service standard. A 6,000 m² building, currently 75% storage, 25% marshalling and dock.

Order history shows the top 30 SKUs are doing 80% of line moves. The next 70 SKUs do 18%. The bottom 100 do 2%. That's a steep curve — exactly the shape cross-docking rewards.

Inbound predictability: the top three suppliers, who between them ship 60% of fast-mover volume, are hitting their delivery windows 92% of the time. The next five are around 80%. Below that it gets ragged.

Customer geography: 70% of routes land in five postcode clusters. Two of those clusters are dense enough that several routes serve them daily.

Volume per drop: the fast-movers go out as mixed pallets — typically four to six SKUs per drop, all from the top 30. The slower SKUs go out as case-picks tagged onto those pallets.

Storage cost: £85 per m² per year, fully loaded. That's mid-range — not cheap, not expensive.

Here's how I'd shape it.

Cross-dock the top 30 SKUs from the top three suppliers. These lines arrive in the morning window, get sorted onto outbound route lanes against the day's order file, and ship out the same afternoon for next-day delivery. Because they all come from reliable suppliers in tight windows, the timing works. Because they're full-pallet or mixed-pallet inbound, the sort is manageable. Because the top three suppliers cover the bulk of fast-mover volume, you don't need every supplier to play — you just need the ones that move the most lines.

Keep the next 70 SKUs in a slim hub-and-spoke operation. Pick locations, replenished from bulk, picked to route. This is your buffer against supplier wobble — when an inbound is late, the picks still happen because the goods are in stock. These 70 SKUs are also where the case-pick complexity lives, and they're better served by a stable pick face than by a chaotic cross-dock floor.

Deflict the bottom 100 SKUs. They're 2% of moves. Some are obsolete. Some are seasonal stragglers. Some belong to one customer and could be billed as dedicated stock at a premium. Run the conversation: each SKU either justifies its storage cost or it doesn't.

The physical building changes too. You don't need 75% of the floor as storage anymore — closer to 50% would do. The reclaimed 25% becomes a cross-dock lane with route bays, a sort area, and enough dock doors to handle the morning inbound peak. Roughly: 12 inbound doors active between 6am and 11am, 18 outbound doors active between 12pm and 4pm, with the sort happening in between.

Labour shape changes too. The afternoon pick team gets smaller. The morning cross-dock and sort team gets bigger. Total headcount typically comes down by 10 to 15%, but the people you keep are more highly utilised, and the labour-per-line on the fast-movers drops by close to half.

The single biggest risk in this operation is the morning. If one of the top three suppliers misses their window, you lose cross-dock volume and have to pick from stock you no longer hold. Mitigations: keep a two-day safety stock on the absolute top 10 SKUs in a dedicated rack, treat that rack as the contingency lane, and review supplier performance weekly. Three months in, you'll know which suppliers can really handle the discipline and which can't, and the cross-dock SKU list will adjust.

That's the move for this shape of business. Not every 50-route op will look like this. If the SKU curve were flatter, if the suppliers were less reliable, if the customers were spread across 50 postcodes instead of 5 — the answer might be "stay in hub, optimise the picks". The framework is the same: thresholds, six questions, trigger conditions, and a SKU-level plan.

Where the software earns its keep

Cross-docking and hybrid only work when the order file, the inbound schedule, and the route plan are all in the same system. The moment those live in three different spreadsheets, the sort breaks down and the cross-dock lane turns back into a storage area within a fortnight.

A network designed this way needs to know, by 9am each day, which inbound pallets are landing into cross-dock and which into storage, which SKUs are on which routes, and how to sequence the sort so the trucks leave loaded in route-stop order. That's what good distribution optimisation tooling is for — not just route planning, but the inbound-to-outbound matching that makes hybrid networks viable.

If you're working through a network redesign for a regional distribution centre, the question worth sitting with isn't "should we cross-dock?" It's "which SKUs from which suppliers can we cross-dock — and what does the building need to look like to make that work?" Get the shape of the answer right and the operational decisions fall out of it.