Future process: how to evaluate measures

A future process, often called the to-be or target process, is the current process with a set of decided changes applied. Each change is a measure with an owner, a due date, an investment and a concrete effect on the model. Its value is the difference in the metrics with and without it; payback is the investment divided by the monthly saving.

From problem to measure

A measure answers a problem at a step (see finding process weaknesses). To be useful later, it needs more than a sentence. In the app a measure carries:

• a text that says what will be changed,
• an owner and a due date, so that it is clear who does it by when (the checks remind you if either is missing),
• a status: open, in progress or done,
• a one-off investment in the process currency,
• its changes to the model, and a switch that includes or excludes it from the evaluation.

Effects as changes to the model

The core idea: a measure does not get a guessed saving typed into a field. It changes the model, and the calculation tells you what follows. The possible changes are:

• a new processing time, waiting time, query effort or query delay at a step,
• a new query rate or new material cost at a step,
• a step becomes automated, or moves to another IT system,
• a step is removed (its predecessors are connected to its successors),
• a new share at a decision exit,
• a new handoff medium on a flow,
• a new number of people, availability or cost rate for a lane.

Effect and payback

The effect of one measure is the set of metrics with this measure minus the metrics without it, both on the current model. The effect of all enabled measures together is calculated the same way. Each measure can be switched on and off, so you see its own contribution.

The saving per year is the difference in annual cost. Payback is only given when that saving is positive.

Worked example: order number mandatory

In the worked example “Invoice check” (1,000 invoices a month), one invoice in five is incomplete, usually because the order number is missing. Each incomplete invoice goes back to the supplier and waits five working days (2,400 minutes) before it is checked again. The measure “Remind suppliers in writing that the order number is mandatory” costs €3,000 and is expected to halve the share of incomplete invoices from 20% to 10%. In the model this is two changes: the exit “no” of the decision “Complete?” gets 10%, the exit “yes” 90%.

Reading the result

The saving per case is small, only €0.51, because a query costs little processing time: the check is repeated, which is 4 minutes. With 12,000 invoices a year it still adds up to €6,111, so the investment of €3,000 pays back in 3,000 ÷ (6,111 ÷ 12) = 5.9 months.

The larger effect is on lead time: a third of a working day less on average, because fewer invoices wait five days for the supplier. And accounts payable drops from 101% to 97.5% utilization, so the backlog of 10 invoices a month disappears. A measure that looks modest in euros can be the one that makes the process stable. That is why the comparison shows all metrics, not just cost (see process cost calculation and staffing calculation).

Deriving and comparing the future state

When the measures are decided, the app turns the current process with all enabled measures into a separate future process. It has the same lanes and steps with the changed values; removed steps are gone, and the measures stay as documentation, marked as applied, so their effect does not count twice. The current state is kept as the comparison baseline. The future process can then be edited further in the same notation, for example to redraw a step that a new system will take over. The future-state template for incoming invoice processing was created exactly this way from the current one.

The comparison places the current and the future process side by side on the same metrics, with the difference: lead time, processing and waiting time, cost per case and per year, handoffs, media breaks, queries and loops, share without queries, full-time equivalents and highest utilization. That is the basis for a decision paper or a steering meeting.

New processes: plan backwards

Not every future process starts from a current one. For a process that does not exist yet, start at the result and ask “what must have happened before?” until you reach the trigger. Then add volumes and estimated times as for a current process (see volumes and case counts) and check whether the teams have the capacity.

Frequently asked questions

Keep learning