A multi tier storage system in Singapore earns its cost through pick faces per square metre, not through storage volume. CBRE put prime logistics occupancy at 95.8% as at Q1 2026, so operators are building up inside leases they already hold. This blog walks you through tier economics, clear height, pick path and throughput, starting with rack supported picking decks.

What is a multi tier storage system, and how is it different from a mezzanine?

A multi tier storage system is shelving or racking built two or three levels high with walkways at each level, where the storage structure itself carries the deck. A storage mezzanine is a platform you put things on. A multi tier system is storage that happens to have a floor. The distinction shows up in what each one is optimised for. A mezzanine maximises square metres of usable floor. A pick module maximises SKU faces reachable on foot, which is a different objective and produces a different structure. Tier decks are usually open grating or steel panel rather than solid decking, walkway widths run tighter than a mezzanine’s, and vertical circulation is designed around order flow instead of around forklift access. Longspan and boltless shelving are the usual building blocks, and the longspan shelving specifications page sets out the standard bay sizes those modules are assembled from. Worth noting: the terminology is inconsistent across suppliers. Some quote a two tier shelving system, some a multi level picking module, and the same drawing appears under both names. Our note on boltless versus shelving terminology covers the same confusion one level down.

When does a Singapore fulfilment operation actually need a second tier?

Order profile decides it, not SKU count. The two numbers that matter are lines per order and picks per SKU per day, and a warehouse holding 8,000 slow moving SKUs has a stronger case for a tier than one holding 12,000 fast moving ones. The reasoning is straightforward. Slow movers consume pick face without generating picks, so they push your aisle run longer without adding throughput. Lifting them to a second tier recovers ground level face for the SKUs that actually move, which shortens the pick path where the volume is. The case gets stronger when demand is growing but the lease is not. Singapore’s retail ecommerce market was S$5.8 billion in 2022 and CBRE projected it reaching S$9.2 billion by 2027 at a 9.9% compound rate, which is line growth landing on floor plates that are not expanding. Where this breaks down is a flat demand curve on fast moving stock. A tier does not make a picker faster. It gives you somewhere to put the SKUs that were slowing everyone down.

How much clear height do you need per tier?

Two tiers need roughly 5.5m of internal clear height and three tiers need close to 8m, which is why two tiers is the practical ceiling in most Singapore flatted factory stock. Work the stack from the roof down. Each tier needs about 2.2m of working headroom for pickers, the deck and its supporting steel take 200mm to 300mm, and sprinkler and services clearance sits above the top tier. A unit with 6m internal clear height gives you two comfortable tiers. A unit with 4.5m gives you one tier and a mezzanine argument you should probably lose. Ground level is the tier that gets squeezed. Designers protect the upper working height and let the ground level absorb the deck depth, which produces a bottom tier where an operator with a trolley cannot stand upright at the shelf face. Set the ground level first and let the top tier take the remainder. For units under 5m the honest answer is usually taller single level shelving with a narrower aisle, and our note on low ceiling warehouse options sets out where that line sits.

What does a second tier do to pick path length?

It doubles pick faces per square metre of floor and does nothing to walking distance unless orders are zoned by tier. This is the calculation most tier proposals skip. Run the numbers on a single aisle. A 30m run with 2.0m bays double sided gives 30 bays, and at 5 shelf levels with 4 SKU locations per shelf that is 600 pick faces along 30m of walking, or 20 pick faces per metre of aisle. Add a second tier over the same footprint and you hold 1,200 faces on the same ground area. The picker still walks. If a typical order pulls three lines from the ground tier and two from the upper one, you have added a staircase or a lift trip to every order, and the tier has made the pick slower rather than faster. The design that works assigns slow movers upstairs, keeps fast movers at ground level, and consolidates at a single downstream point so most orders never cross tiers. Zone your slotting before you build the deck, because a tier built without a slotting rule is a taller version of the same problem.

Where does the throughput bottleneck move?

To vertical transfer. Tiers multiply pick faces without multiplying the ways product gets down to despatch, so the conveyor drop zone, goods lift or staircase becomes the constraint the day the module goes live. Three transfer methods cover most Singapore fulfilment builds. A gravity or powered conveyor from each tier into a consolidation lane, a goods lift serving tote trolleys, and manual carry down a staircase. The third is the cheapest to build and the first to fail at volume, because a picker carrying totes downstairs is not picking. Size the transfer against peak hour lines, not average daily lines. A module handling 400 lines an hour at peak with an average of six units per tote needs roughly 67 tote movements an hour off the upper tier, and a single staircase carry cycle at two minutes round trip supports 30. Design the drop zone before the deck layout. It sets where the staircase lands, which sets the aisle orientation, which sets the bay grid. Working the other way round produces a module where the conveyor has nowhere sensible to go.

What deck and walkway specification should you write?

Open steel grating for the walkways, mesh or solid panel infill at the shelf face, and edge protection with toe boards on every open side. Grating is the default because it lets sprinkler discharge, light and air through the deck, which matters more on a picking module than on a storage platform. Walkway width is the operational decision. A single picker with a trolley needs about 1.2m of clear walkway, and two way trolley traffic on the same tier needs closer to 1.8m. Specifying 900mm to save steel produces a tier where pickers wait for each other, which is the same throughput problem in a different place. Fall protection is not optional and is where cheap quotations show. Every open edge needs a handrail, a mid rail and a toe board, and pallet gate positions need a controlled opening rather than a removable chain. Write the walkway width, edge protection detail and deck material into the tender rather than accepting a supplier standard. Two quotations that look similar on price separate immediately once those three lines are specified.

Do SCDF and BCA treat a pick module as a mezzanine?

Yes. A tier deck that people walk on and goods sit on is a mezzanine floor regardless of whether the supplier calls it shelving, and both agencies assess it that way. BCA treats it as building work under the Building Control Act, so structural plans go in through a Qualified Person who is a registered Professional Engineer. SCDF assesses it under Fire Code 2023, where Clause 9.6.1a.(2) permits a single open staircase only where six conditions are met together, including aggregate accessible floor area not exceeding 60m2 per unit and use restricted to “store and/or ancillary office”. That 60m2 figure catches almost every real pick module, which means most multi tier builds carry two remotely located exits by design. Budget for the second staircase from the first drawing rather than discovering it at submission. The use restriction matters as much as the area. Packing benches and value added services on a tier deck change what the space is used for, and that is a conversation to have with the Professional Engineer before the layout is fixed, not after.

What changes for a 3PL holding multiple clients?

Contract length changes the structure specification. A 3PL running two year client contracts should build tier modules from bolted components that come apart and reconfigure, because the client mix will change before the steel wears out. Segregation is the second difference. Client separation on a single level is a floor marking exercise. Across tiers it becomes a slotting and access question, and the cleanest arrangement puts one client per tier zone rather than splitting each client vertically, which keeps stock counts and cycle audits contained to one level. The commercial argument follows from that. A tier module that can be reconfigured in a weekend lets you win a client whose SKU profile does not match the current layout, and that flexibility is worth more to a 3PL than the marginal steel saving from a welded structure. Our multi client layout planning guide covers the single level version of the same problem. CBRE’s Asia Pacific Logistics Occupier Survey found respondents ranking “proximity to markets and customers” among the most important site criteria, which is precisely why 3PLs intensify existing well located units rather than moving to cheaper ones further out.

How do you choose between a tier, taller shelving and more space?

Match the answer to what is actually short. Short on pick face with height available means a tier. Short on storage volume with pick face adequate means taller shelving and a narrower aisle. Short on both with no height means more space. Two tests separate the first two cases. Count how many SKUs currently have no dedicated face and are stored in reserve locations, and measure how far a picker walks per line on a typical order. High reserve counts with long walks points at a tier. Low reserve counts with short walks points at storage height. The financial test comes last, not first. A tier module sized at 300m2 of deck buys back that floor area for a fraction of what the equivalent leased space costs at current rents, and the payback period worked example shows how to run that arithmetic on your own numbers. Moving remains the right answer when the constraint is dock doors, floor loading or turning circles. No amount of vertical intensification fixes a building with the wrong ground floor, and the expansion options compared guide sets out that decision in full.

The metric that decides it

Multi tier storage is a pick face decision wearing a storage costume. The tier pays when slow moving SKUs consume ground level face and orders can be zoned so most picks stay on one level. It costs you when orders routinely span tiers and every line picks up a staircase trip. Settle the slotting rule and the vertical transfer method before the bay grid. Both set the layout rather than following it. Send us your SKU count, lines per order and internal clear height, and we will tell you whether a second tier earns its floor.

Frequently asked questions

How many tiers can you build in a Singapore warehouse unit?

Two in most cases. Each tier needs roughly 2.2m of working headroom plus 200mm to 300mm of deck and supporting steel, so three tiers requires close to 8m of internal clear height. Flatted factory units in Singapore typically run 4.5m to 6m, which supports one or two tiers.

Can you run a conveyor between tiers on a multi tier module?

Yes, and it is the transfer method that scales. Gravity chutes and powered take away conveyors from each tier into a consolidation lane remove the manual carry that limits staircase served modules to roughly 30 tote movements an hour. The conveyor position has to be fixed before the deck layout, not after.

Is a multi tier system worth it below 2,000 SKUs?

Rarely, unless the SKUs are bulky. Below that count a single level of longspan shelving with a tighter aisle usually delivers the same pick face at lower cost and without triggering a BCA structural submission. The threshold is pick face shortage, not SKU count on its own.

Do pickers need different training on a multi tier module?

Yes. Working at height changes the risk profile, and pickers need briefing on edge protection, pallet gate discipline and load limits on the deck. Singapore’s Workplace Safety and Health Act places that duty on the occupier, and the tier deck carries its own imposed load rating that staff have to know.

Can a multi tier module be dismantled when a 3PL contract ends?

Yes, if it was specified with bolted connections and mechanical anchors. Modular shelving based tier systems come apart into their original components and reconfigure into a different bay grid. Welded structures with drilled and grouted base plates do not, which is why contract length belongs in the specification.