Evaluating automation: Useful output, rework and ongoing costs

Automation is worthwhile for your business when it produces useful work with a reasonable total effort. Counting how many workflows started does not answer that question. An output generated quickly may need a long review afterwards. Conversely, a small step performed reliably can release capacity exactly where your team needs it. Before expanding an automated process, compare both approaches through to the point where the result is genuinely ready for use.

This article develops a working calculation for three workload levels. You first establish the previous effort, add review and ongoing supervision, and then examine the economic difference. Every number is an explicitly invented assumption for a fictional business. None describes webRichtung prices, performance or success rates. The worksheet is a decision aid you maintain yourself. It does not imply that Cockpit provides an automatic financial or return on investment dashboard.

Measure the baseline through to a completed work step

Define where the comparison starts and ends. In our example, a case begins once all required information is available. It ends when an internally usable summary has been checked and incorporated into the intended workflow. Simply generating a text file would not reach the same endpoint. This boundary prevents a common mismatch: counting every activity in the manual process while measuring only the fast first draft in the automated one.

Observe several ordinary cases alongside the exceptions that actually occur. Record active handling time, questions and corrections. Waiting for a reply affects elapsed time, but it is not automatically working time. If an employee handles other tasks while waiting, those hours cannot all be treated as saved personnel costs. Keep the two measures separate so that a shorter turnaround and a lighter workload do not become interchangeable claims.

Check the previous quality as well. If the manual process regularly misses information, a more consistent workflow may offer another benefit. Describe it concretely first, such as a more complete handover. Do not assign a financial value without supporting observations. An honest time comparison with a documented quality standard is more useful for an initial decision than a large benefit figure whose components nobody can explain or verify.

Explore Cockpit for an overview of your platform use and get started on the module page.

Count usable results and all supervision time

Start measuring the new process at the same point. Working time includes preparation, input, review, correction and adoption of the result. If a case requires a second run, keep that effort attached to the case. If the output is discarded and recreated manually, it must remain in the evaluation. Otherwise you measure only successful outcomes and understate the real effort involved in operating the process.

Separate work on individual cases from ongoing supervision. A brief check of every draft belongs to each case. Reviewing the instructions monthly, investigating an interruption or discussing changed input data may take time independently of volume. That supervision does not automatically disappear when usage falls. This is precisely why the same workflow can call for different decisions at a low workload and a high one, even if each individual output behaves similarly.

Count unique cases that have become ready to use. Repeated technical executions may incur costs, but they are not additional business results. If ten cases require twelve executions, the outcome count remains ten. Include the cost of repeated runs in actual operating costs. Also establish whether the review really happened. Shorter handling achieved by skipping checks is not a fair comparison when the expected quality standard is supposed to remain the same.

Build a worksheet with visible assumptions

Our fictional business assumes twelve minutes of active work per case in the previous process. The automated workflow still requires an average of five minutes, including normal review and rework. It also needs three hours of supervision each month. For planning purposes, the company values an internal working hour at €40. That value represents capacity; it is not automatically an amount that will disappear from the business’s future cash payments.

For this model only, we also choose external running costs of €0.40 per case and €60 of monthly external expenditure independent of volume. These are invented calculation inputs, not supplier prices or descriptions of a tariff. Setup and familiarisation are assumed to cost €600 once. Replace every input in your own worksheet with documented effort and actual applicable terms. The variable amount must also account for any repeated runs that genuinely incur charges.

The baseline calculation is monthly cases multiplied by twelve minutes, divided by sixty, and multiplied by the internal hourly value. For the new process, use five minutes per case, add three supervision hours, and value the total at the same €40 per hour. Then add the assumed variable and fixed external costs. The difference is the monthly valued benefit before the one-off setup effort. Keeping these components visible makes later changes easier to explain.

Work through three workload levels completely

At 50 cases a month, the previous process requires ten working hours, valued at €400. The new process needs 250 minutes of case work plus 180 minutes of supervision, giving 430 minutes in total. That is seven hours and ten minutes, valued at €286.67. Add €20 of variable costs and €60 of fixed external costs. The resulting total is €366.67 against the previous €400, leaving a valued monthly benefit of €33.33.

At 200 cases, the baseline is 40 hours valued at €1,600. The new process takes 1,000 minutes plus 180 minutes of supervision: 1,180 minutes, or 19 hours and 40 minutes. The assigned labour value is €786.67. Variable costs rise to €80, while the fixed external assumption stays at €60. The new total is therefore €926.67, and the monthly difference is €673.33. Both approaches still end with the same usable result.

At 400 cases, the baseline comprises 80 hours valued at €3,200. The new workflow requires 2,000 minutes plus 180 minutes, giving 2,180 minutes or 36 hours and 20 minutes. Its labour value is €1,453.33. With €160 of variable costs and €60 of fixed external expenditure, the total becomes €1,673.33. The valued monthly benefit is €1,526.67. All monetary figures are rounded to two decimal places only for presentation, rather than rounding the working time first.

The calculation demonstrates why volume belongs in the decision. Three supervision hours and the fixed external amount are spread across more cases at higher utilisation. It does not establish that 400 cases actually exist or need the same average rework. Only use higher volumes if they match real business demand. Running additional processes that serve no business purpose merely to keep an automation busy creates no economic benefit, regardless of how impressive the activity count appears.

Distinguish released capacity from actual savings

In the middle scenario, 40 previous hours become 19 hours and 40 minutes of internal work. This releases 20 hours and 20 minutes of capacity. The valued time difference is €813.33; subtracting the €140 of external model costs leaves the stated €673.33. That distinction matters. The company does not necessarily have €673.33 more cash. If salaries stay unchanged, it may initially incur additional external expenditure while its personnel payments remain the same.

Decide how the available time will be used. Are there delayed quotations, realistically scheduled customer jobs or outside work whose cost can actually be avoided? Describe a specific use and name a responsible person. If the team uses the time to improve availability, that can be worthwhile without immediately claiming additional revenue. A cash-based calculation should include only expenditure that genuinely disappears and additional contributions supported by a credible basis for expecting them.

Do not count the same benefit twice. If released hours already appear as economic value, do not simply add all the revenue those hours might produce. Revenue is also different from the contribution remaining after the costs of delivering that work. This article deliberately keeps the comparison at the value of working effort. You can add a separate capacity decision later, once specific follow-on work, delivery requirements and associated costs have been established.

Test rework and one-off effort deliberately

The five minutes remaining in the new process are the example’s most sensitive assumption. If they rise to eight minutes at 200 cases, case work takes 1,600 minutes, plus 180 minutes of supervision. The total is 29 hours and 40 minutes, with a labour value of €1,186.67. Adding the unchanged €140 of external costs gives €1,326.67. Compared with the previous €1,600, the monthly benefit falls to €273.33. The workflow still saves valued effort, but the cushion is much smaller.

The model therefore remains positive while leaving considerably less room for setup effort and uncertainty. Measure rework on cases of differing difficulty. An average based on unusually clean inputs may misrepresent everyday operation. Also check whether supervision really stays at three hours as volume grows. If more cases introduce new types of exceptions, update the supervision assumption before using the calculation to justify expansion. Scaling an input should not silently freeze every other cost.

The one-off €600 assumption must be recovered as well. With the original model inputs unchanged, simple calculated payback would be approximately 18 months in the smallest scenario, less than a month in the middle one and less than a month in the largest one. This offsets valued benefits against setup effort; it is neither a cash forecast nor a promise. What matters for your decision is whether the underlying volume and output quality remain plausible throughout the period considered.

Decide whether to expand, adjust or stop

For the small scenario of 50 cases, a €33.33 monthly benefit is a thin margin. Additional supervision could absorb it quickly. A reasonable decision could be to simplify setup or restrict the workflow to a clearly defined bottleneck before investing further effort. The low figure does not mean every automation is unsuitable at low volume. It shows the consequences of these particular assumptions for this particular work step. A different workflow requires its own evidence.

In the middle scenario, limited expansion may make sense if everyday observation confirms five minutes including rework and the released time is actually needed. Decide beforehand when to review the result, perhaps after one fully observed operating month. The decision should state who owns supervision and which changes trigger a new calculation. This turns a one-off trial into a controlled business decision, with a clear point at which the original assumptions can be challenged.

Stop or substantially change the workflow if it fails the required quality standard or works only through persistently excessive rework. A positive time balance cannot make an unusable output useful. In that case, secure outstanding cases and maintain the previous working route. Already incurred setup costs remain documented, but they are not a reason to accept more poor results. Future effort should be judged by what it can achieve from now on, rather than by a desire to justify the past.

Use Cockpit as a starting point for ongoing observation

The webRichtung cockpit product page describes metrics and events from your platform use. Cockpit reads this information without intervening itself. That overview can support regular attention to how the platform is being used. Check the actual product offering to establish which values are available for your workflow. Internal working time, valued capacity and the economic calculation developed here are additional elements of your own business evaluation, rather than promised financial product features.

Start with a manageable worksheet and someone responsible for maintaining it. Record the period, counted cases, quality standard and cost source together. After the first complete period, compare the assumptions with the observed workflow. If a figure differs, update the calculation and explain the difference. You do not need a large collection of indicators. You need a traceable connection between useful work, necessary supervision and the decision you intend to make next.

On the Cockpit page, you can learn about the intended overview and create an account through the existing starting route. Bring a specific workflow whose expansion you are considering. A clearly recorded baseline and a complete observed operating period make it easier to judge where additional automation can help your company. They also reveal where a clearer task definition or less rework needs to come first, before more volume is introduced.

Start with Cockpit and establish a clear basis for deciding your next automation step.