Warehouse management:custom WMS or SaaS?

What a WMS has to do

A warehouse management system (WMS) controls all movements in the warehouse. According to the common classification, which is based on the German guideline VDI 3601, this includes:^[1]

• Master data and warehouse structure: items, storage locations, zones.
• Inventory management: quantities per location, reservations, stocktaking.
• Goods receipt: advance notice, receiving, quality inspection, putaway.
• Internal processes: relocation and replenishment.
• Goods issue: picking, packing, shipping.

In practice, further requirements are added depending on the industry, such as batches and best-before dates for food, serial numbers or different picking methods. These are exactly the points where standard solutions differ most.

What warehouse SaaS costs today

The following overview shows the entry prices of some solutions that are widely used in German-speaking countries. All prices are net and in the vendor’s currency.

Vendor information retrieved on 16.09.2026. Ranges reflect monthly or annual billing. Prices in Swiss francs and set-up costs are not included.

The entry prices look low. What matters is how they grow with your volume: fees per order, per device or per user add up quickly with thousands of orders a month. Set-up, integrations and add-on modules often come on top.

Why SaaS costs keep rising

Several analyses show that software subscriptions are rising in price much faster than general inflation. Vertice, a provider of SaaS purchasing software, measured SaaS inflation of 8.7 % for 2023, 11.4 % in January 2025 and 16.4 % in June 2026.^{[7, 8, 9]} According to Vertice, 33 % of the contracts analysed contained uncapped price increase clauses.^[8] Since Vertice itself sells software in this market, these figures should be read with care.

Gartner expects global software spending to grow by 15.2 % in 2026 and names generative AI as one driver of higher software prices.^[10] Zylo’s 2026 SaaS Management Index also shows a median SaaS spend of USD 9,455 per employee, with 36 % of licences unused.^[11] Both studies mainly cover larger US companies, but the trend is relevant for SMEs too.

How AI is changing development

For a long time, custom software was too expensive for SMEs. AI tools are changing that calculation, though less clearly than often claimed:

• In a controlled experiment by GitHub and Microsoft, developers completed a single programming task 55.8 % faster with Copilot.^[14]
• McKinsey observed that documentation and code generation took about half as long, but for complex tasks the time saved was below 10 %.^[15]
• A 2025 METR study with experienced open-source developers even found that they took 19 % longer with AI. A 2026 update points to a speed-up, but it is not statistically significant.^{[16, 17]}
• Google’s 2025 DORA report concludes that AI amplifies a team’s existing strengths and weaknesses: more throughput, but lower delivery stability.^[18]

For your decision, this means routine work gets cheaper, while careful architecture, testing and quality assurance remain necessary. There is no reliable figure for how much cheaper an entire project becomes.

Risks of custom development and how to limit them

Large IT projects have a poor reputation, and not without reason. A study by McKinsey and the University of Oxford of more than 5,400 projects with budgets above USD 15 million found an average budget overrun of 45 %.^[13] The main lesson for SME projects: start small and deliver in steps.

• Dependence on the developer: source code in your own repository, complete documentation and clear rights to the code. In addition, the source code can be deposited with a neutral escrow agent.^[21]
• Data protection: if a service provider processes data for you, you remain responsible. Under the Swiss Data Protection Act, you need a data processing agreement covering instructions, data security and rules on subcontractors.^[22] The Act also requires privacy by design and by default.^[23] Violations can lead to fines of up to CHF 250,000 against the responsible individuals.^[24]
• Operation: hosting, backups, security updates and monitoring must be covered by contract.
• Scope: do not build everything yourself. Gartner recommends a mix of buying, building and integrating, with clear rules on when each applies.^[12]

KPIs to measure a WMS against

A study by the Technical University of Munich found an error rate of around 0.3 % of order lines when picking with paper lists. An error cost around EUR 25 when found internally and around EUR 210 when the customer found it.^[19] According to the WERC benchmarks for 2025, the best warehouses achieve a picking accuracy of at least 99.68 %.^[20]

Useful KPIs for your decision are therefore picking accuracy, time from goods receipt to putaway, lead time per order, inventory discrepancies and software cost per order.

Decision guide

A middle way often makes sense: standard software for accounting and the shop, and a lean custom warehouse solution for the processes that make your business different.

Frequently asked questions