Goods-to-person automated warehouse: when it really pays off, with numbers and thresholds

In almost every traditional warehouse the biggest cost item in picking is a single one: the time operators spend walking. That is where to start when evaluating a goods-to-person system, because it is exactly the item automation goes after, and because the cost of manual picking can be measured well, with data the company already holds. Order lines per day, number of items and turnover, labor cost, space, shifts: these numbers, more than the technology, say whether a warehouse is a serious candidate, before anyone talks to a supplier. And the same numbers, read honestly, also say when goods-to-person isn't needed.

Why goods-to-person changes the picking equation

In a traditional warehouse the operator goes to the goods: walks the aisles, finds the location, picks, comes back. In a goods-to-person system the flow is reversed. Mobile robots lift shelves and bring them to picking stations, where the operator stays put and picks with software guidance. Wurman, D'Andrea and Mountz described this architecture in AI Magazine in 2008, in a paper that has remained a reference for the field. The reason it matters lies in the reference data of warehouse logistics. Tompkins and colleagues, as cited in the review by de Koster, Le-Duc and Roodbergen in the European Journal of Operational Research (2007), estimate that picking can account for up to 55% of a warehouse's operating cost. In traditional picking, the same source adds, travel takes about half of the operator's time.

Goods-to-person removes almost all of that item, which is why productivity per operator can rise sharply without asking anyone to move faster. The mobile-shelf manufacturer that DIPITECH represents states picking productivity up to three times the manual figure. It is an interested party's number, to be measured on your own line profile, and the goods-to-person automated warehouse page reports the conditions under which it is stated.

What walking costs you today

Every serious evaluation starts from the current cost of picking, and the formula only uses internal data. Annual picking cost is the number of dedicated operators multiplied by the hourly company cost and by the hours worked in a year. The walking share is that cost multiplied by the percentage of travel time, which without your own measurements can be estimated at around half, as in the literature just cited.

An example, declared as such. A warehouse with 6 operators dedicated to picking, on one 8-hour shift for 220 working days, uses 10,560 labor hours a year. If about half is travel, the company pays every year for more than 5,000 hours of walking, the equivalent of two or three full-time staff whose job consists of walking aisles. That is the quantity goods-to-person turns into picks: the same volume of lines is handled with fewer hours, or growing volume is absorbed without looking for people who can't be found. Picking errors come on top, each with its return and reshipment, and display-guided picking at the station reduces them.

Payback thresholds: order lines per day

The thresholds below are our own order of magnitude, derived from two written assumptions. The first is manual productivity of 65 lines per hour per operator, the same used in the example of our guide on choosing an automated warehouse, and the second is walking equal to half of picking time. At 150 lines a day picking takes just over two labor hours, and walking just over one: even removing all of it, the annual saving stays small. At 1,000 lines a day picking hours rise to about 15 and walking alone takes almost eight, a whole person every day. Between these two ends, the other variables decide.

Lines alone are not enough, because the number of items and the turnover curve matter too. The profile that suits goods-to-person is a range from hundreds to tens of thousands of small and medium items with spread-out turnover, typical of e-commerce, spare parts, cosmetics, electronics and pharmaceuticals. With very few fast movers the right goods are always in front of the operator already, and a well-designed flow rack may be enough. With only slow movers there are few walking hours to recover in absolute terms. Goods-to-person pays where the mix forces operators onto long, fragmented routes today.

Space, shifts and seasonality move the numbers

A calculation on picking alone, though, leaves out three multipliers. The first is space: mobile shelves remove most aisles and compact storage, and the manufacturer we represent states up to 30% of floor space recovered. Where the warehouse is full this can become the biggest saving of all, because it postpones an extension or a second site, or frees area for production. The second multiplier is shifts. The investment is spread over operating hours, so a system working two or three shifts earns two or three times the saved hours on the same capital. The third is seasonal demand. In busy periods a manual warehouse hires seasonal staff who have to be trained in a hurry, with worse productivity and more errors exactly when volume peaks. A modular goods-to-person system adapts by adding robots to the fleet instead. That scalability is hard to budget for, but very real in November.

That is why the serious answer to the payback question is sizing on real data, and why it pays to be wary of anyone promising a payback without having seen the data.

From data to payback

Lines, items, floor plan and organization are the basis for sizing the robot fleet, shelves and stations, and payback is calculated with a method: total investment, meaning system, integration and training, divided by the net annual saving. The saving adds up recovered picking hours times hourly cost, the value of freed space, fewer errors and less overtime and seasonal labor, net of the system's annual costs. The full method, with the hour conversion factor and volume sensitivities, is in our guide on choosing an automated warehouse.

A specific advantage of mobile-shelf systems is their installation profile. For standard systems the manufacturer states a physical installation of two to three weeks, possible because the robots follow floor references and no civil works or bolted fixed structures are needed. The overall project, with data analysis, WMS integration, training and fine-tuning, takes longer, and it is the part that decides the result.

The profiles where goods-to-person doesn't pay

The first wrong profile is too few lines. Below 150-200 lines a day the hours saved rarely justify the investment, and it is better to optimize routes, locations and order batching in manual picking. The second concerns oversized or very heavy items: the system works with items that fit the compartments and the load capacity of the mobile shelf, which the manufacturer states at 600 kg. Full pallets, long or heavy items call for robotic forklifts and AMRs for pallet handling or traditional solutions.

The third is a range of very few fast movers. If 90% of lines fall on a few dozen codes, a flow rack or a well-designed pick zone delivers most of the benefit at a fraction of the cost. The fourth is a flow dominated by full pallets in and out, where the issue is material handling and the right evaluation is AMR ROI and TCO. Very often the correct answer is hybrid: goods-to-person for fast-moving small parts, handling vehicles for pallets, conventional shelving for bulky slow movers.

Goods-to-person, conventional shelving and mini-load compared

What to send us for an assessment

To place your warehouse in one of the three zones a few figures are enough, as long as they come from the host system rather than estimates:

• order lines per day, with hourly and seasonal profile, ideally twelve months exported from the host system;
• item master: number of items, case dimensions and weights, ABC turnover curve;
• floor plan with clear heights and constraints such as columns, docks and served areas;
• WMS or ERP in use and integration points;
• organization: operators dedicated to picking, shifts, overtime and seasonal staff in busy periods.

You can send them by email to info@dipitech.com or on WhatsApp. The assessment returns a sizing with the return calculation method spelled out and, when that is the case, the conclusion that goods-to-person does not fit your profile. The available solutions are part of our industrial automation and robotic material handling offering. If the choice is still between transport and storage, the comparison of AGVs and AMRs helps separate the two problems.