Direct answer
A WMS business case compares two cost pictures: what your warehouse costs to run today, including the price of its errors, and what it would cost to run with a WMS, including the price of the system and its adoption. The difference, set against a period you are willing to wait, is the case. Start by costing the errors you already have, because they are real, they are in your own records and they are the part most often left out.
This page is a framework, so it contains no percentages, benchmarks or savings claims. Your warehouse is not an average one. For what drives the system side of the equation, see WMS implementation cost, and for how to tell whether spreadsheets are the problem, see logistics spreadsheets outgrown. The WMS readiness assessment helps you see how prepared the operation is before you price anything.
Current baseline
Describe how the warehouse works today in measurable terms: order lines handled per period, receipts per period, number of SKUs and locations, people by shift, and how stock is tracked. Take these from your order system, receiving records and rota, not from memory. Note the peak periods separately, since a WMS is judged in the busiest week, not the average one.
Also record how work is directed today: paper pick lists, spreadsheets, someone's head. The baseline should let a colleague who does not work in the warehouse understand what is being replaced.
Error costs
List the error types you see: wrong item shipped, wrong quantity, stock not where the system says, receipts recorded late, returns not put back into stock, orders promised on stock that was not there. For each, count how often it happens over a period using complaint logs, credit notes, re-shipments and stock count differences.
Then cost each occurrence from its parts: the time to find and correct it, the carrier cost of a re-shipment, the value of goods lost or written off, and any credit given to the client. In symbols: occurrences per period x cost per occurrence. If a cost such as lost goodwill cannot be measured, list it as unpriced and keep it out of the arithmetic.
Labour touches
Count how many times a person handles information or goods per order line: reading a list, writing a note, searching for a location, re-keying data, checking a pick. Each touch is time. Walk the process with a stopwatch on a sample of orders and record the touches and their duration, including the walking and searching that never appears on a job description.
Then be honest about what a WMS changes. It may remove some touches, such as re-keying and searching, and add others, such as scanning. The comparison must include the new steps, not only the removed ones. A person's time counts as a saving only if it lets you handle more volume with the same people or avoid adding staff.
Space and stock
Consider what inaccurate stock information costs you in space and capital. Safety stock held because nobody trusts the numbers, locations kept empty as a precaution, stock that cannot be found and is reordered, and slow-moving items that occupy good locations all have a cost. Measure what you can: stock-count differences, reorders of items later found on site, and time spent searching.
Be careful here. Improved visibility does not automatically release space or cash. It only does so if you then change purchasing or layout decisions. Include only the effects you will act on, and label the rest as possible upsides.
System cost
List the full cost of the system, not only the licence or build. Include software or development, configuration to your workflows, labels and scanning equipment, master data cleaning, integration with your order or accounting system, training time, the temporary drop in throughput around go-live, and the running cost afterwards, including support and changes as your operation evolves.
Ask each supplier for one written scope stating what is included and excluded, so that the figures can be compared. For the sequence of work that generates these costs, see the warehouse management system page.
Risks
A business case should price its own risks. Common ones are data that is not clean enough to load, staff who resist a new way of working, hardware that does not suit the environment, an integration that takes longer than expected, and a go-live period in a busy season. For each, write what would happen, how likely you think it is, and what you would do about it.
Do not attach precise probabilities you cannot support. A qualitative rating with a named mitigation is more honest than a number. Then repeat the arithmetic with pessimistic values for the largest inputs and see whether the case survives.
Comparison with staying as-is
Staying as you are is an option with a cost, not a free default. Project the baseline forward: if volume grows, what happens to errors, headcount and space? Also include the cheaper alternatives, such as fixing the spreadsheet process, tightening labelling, or improving a few specific workflows, because a WMS should be compared with the best alternative, not with doing nothing.
Put the options side by side using the same cost headings: running cost, error cost, one-off cost and risk. A fair table often shows that the case depends on one or two inputs, which tells you where to measure more.
Decision
Summarise on one page: the baseline, the priced errors, the measured touches, the full system cost, the risks with mitigations, the comparison and the payback threshold you set beforehand. Decide to proceed, to fix specific problems first, to measure more, or to stay as you are.
If it helps to test your framework against a scoped plan, we can review it and give a written scope for the parts of the work we would do, and we will tell you straight if a WMS is not the right next step for your warehouse.
Questions
Why does this page not include savings estimates?+
Because warehouses differ too much for any general figure to be meaningful. The framework shows what to measure so the numbers come from your own records.
Where should I start?+
With the errors you already have. They are documented in credit notes, re-shipments and count differences, and they make the case concrete without depending on forecasts.
Should I include reduced headcount as a benefit?+
Only if you will actually avoid hiring or reassign people to work with a cost effect. Otherwise describe it as capacity gained and keep it out of the totals.
What is the best alternative to compare with?+
Not doing nothing. Compare with fixing the current process, improving labelling and locations, or a smaller step, so the WMS is judged against the best cheaper option.
Is the system cost only the software?+
No. Include data cleaning, labels and equipment, integration, training, the go-live dip and ongoing support. Ask for a written scope stating what is and is not included.