Insights / Guide / Business process reengineering (BPR)
GuideBusiness process reengineering: the complete guide
When a process was built around departments and history, tuning is not enough. How to redesign work end to end, with controls built in, before you automate.
Sometimes improving a process is not enough, because the process was built around departments and history rather than around the customer or the result. Each team has optimised its own part, but the work still crawls from desk to desk, waiting for approvals and being re-keyed into different systems. Shaving minutes off individual steps does little when most of the time is lost between them.
Business process reengineering redesigns the flow end to end, so work moves with fewer handoffs, fewer approvals and less rework. This guide draws on the briefings in our reengineering series. It explains when redesign is the right answer, how to approach it, what it looks like in three common processes and how to get from diagnosis to a tested design in about a month.
Start with the outcome
Agree what the redesigned process must achieve and how you will measure it before anyone draws a new flow. Outcomes should be specific and meaningful to the business: orders shipped within two days of receipt, supplier invoices paid on time without duplicates, month-end close completed in five working days. Vague goals such as “more efficient” give the design team nothing to test against.
Measure the current performance against the same outcomes, so that improvement can be demonstrated. Where data is thin, a short sample of recent cases tracked by hand is usually enough to establish a credible baseline.
Redesign before you automate
Automating a poor process only makes the problem faster. The temptation to buy a system and let it impose a better process is understandable, but systems faithfully reproduce whatever process they are configured to follow, including its unnecessary steps. Redesign first, automate second.
- Question every step. Ask what value it adds for the customer or the business, and remove steps that exist only for historical reasons.
- Combine handoffs. Fewer handoffs mean fewer delays and errors; where possible, let one person or team complete a whole unit of work.
- Design around the output. Organise the flow around the customer order, the invoice or the product, not around departmental boundaries.
- Capture information once, at the source, and let it flow to every step that needs it.
- Then apply technology to the clean, stable process.
One owner, end to end
End-to-end processes need one accountable owner, not four department heads each responsible for their own segment. Without that owner, redesign stalls at the first boundary between teams.
Design controls into the flow
In many processes, control is a separate step bolted on after something went wrong: an extra approval, a second check, a reconciliation at month end. Each adds delay, and because they sit outside the main flow, they are easily skipped under pressure.
A better approach builds approvals, checks and audit trails into the flow itself. Validation happens as data is entered rather than days later. Approval thresholds are tiered so that small, low-risk items pass quickly and attention goes to the large or unusual ones. Systems record who did what automatically, so the audit trail is a by-product of doing the work rather than an extra task.
There is a balance to strike. Removing a control entirely may save time but expose the business to error or fraud, and auditors will rightly ask why it went. The aim is not fewer controls for their own sake but better-placed ones: checks that happen automatically, at the moment the risk arises, rather than manual reviews long after the fact. Involve finance and internal audit early so the new design is agreed rather than challenged later.
Where redesign pays back: three common processes
Order to cash
The path from customer order to cash collected touches sales, operations and finance, and small gaps add up. Errors entered at the quotation or order stage travel through every later step, so validate early: check prices, terms and customer details as the order is captured. Link delivery to invoicing so that bills are triggered by delivery events rather than waiting for paperwork. And send reminders before due dates; they reduce late payment more than chasing afterwards.
Procure to pay
Purchasing needs control, but too many approvals slow the business and encourage workarounds such as splitting orders to stay under limits. Tier the approvals: light for small purchases, stricter for large ones. Match order, goods receipt and invoice before paying to avoid errors and duplicates. Use preferred suppliers with catalogues and agreed terms to reduce effort and improve pricing.
Plan to produce
Production plans often fail because information moves slowly between planners and the shop floor. Planners, buyers and supervisors should all see the same schedule. Record what was made and consumed close to when it happens, not at the end of the shift. Compare plan with actual every week and ask why they differ; the answers point directly to the next improvement.
Define the system needs and the business case
Once the new flow is designed, specify what your systems must do to support it and what must change in them. This turns the redesign into clear, testable requirements for configuration, integration or automation, and it prevents a system project from quietly reintroducing the old process.
Quantify the benefits and costs honestly: time saved, errors avoided, cash released, and the cost of system changes, training and transition. Sequence rollout, training and change so that the organisation is not asked to absorb everything at once.
A four-week plan from diagnosis to redesign
Redesign need not take months. A tight plan keeps momentum and keeps people engaged.
- Week one: understand. Map the current process with the people who do the work and collect basic data on volumes, cycle times and errors.
- Week two: find causes. Analyse where time is lost and errors begin, and rank the causes by impact.
- Weeks three and four: design and test. Design the new flow with the team and pilot it on a small set of real cases, adjusting as you go.
Pilot first, in one site, team or product line, before rolling out widely. A pilot shows whether the design works in practice, surfaces problems while they are cheap to fix and gives you evidence to persuade the sceptics.
Agree in advance what the pilot must show before wider rollout, using the outcome measures set at the start. If the pilot falls short, adjust the design and run it again rather than pressing ahead on hope.
Bring people with you
People will resist loss of control. Redesign often removes approvals that managers valued, combines roles or changes who owns a decision. Involve those who do the work from the start, explain what they gain, such as less chasing, fewer errors and more interesting work, and be honest about what changes for them.
The new process was obviously better on paper. It only stuck because the people who ran the old one helped design it and could explain why to everyone else.
Operations director, distribution client
Done well, reengineering delivers shorter cycle times, fewer errors and less rework, lower cost per transaction and a system-ready design. Choose one end-to-end process that matters to customers or cash, name a single owner, and run the four-week plan on it. The results will tell you whether to extend the approach to the next process.
Source: Intelliversal analysis, drawing on the briefings in this series.
