Business process optimization: a step-by-step playbook

Why optimization is mission-critical for the business

Almost every company eventually hits a ceiling on operational efficiency — somewhere around 60-70% — and stalls there for years. In practice that means 30-40% of your productivity is simply slipping through your fingers. And the justifications always sound reasonable: "it kind of works," "people are used to it," "now's not the time." But the outcome is always the same: margins flatline, there's nowhere left to grow, and meanwhile the competition pulls ahead.

Over the past two years we've helped dozens of companies lift their core processes by 25-40%. Below is the very playbook we use ourselves — one that maps onto almost any business.

Stage 1: Diagnostics and problem identification

Step 1: Choose the processes to analyze

Don't try to fix everything at once — you'll only spread yourself thin. Take 3-5 core processes that consume the most resources:

• For e-commerce: the order process (order → fulfillment → dispatch → delivery)
• For consulting: the sales process (lead → meeting → proposal → contract), the delivery process (contract → execution → report)
• For manufacturing: materials procurement, production, quality control
• For SaaS: customer onboarding, customer support

Step 2: Process mapping (AS-IS)

For each chosen process, document: the start, the end, EVERY intermediate step, the people involved, and the timing.

Example: materials procurement in manufacturing

1. A production worker notices that a material (raw input) is running low
2. They fill out a procurement form (in Excel or the CRM) — 30 minutes
3. The form goes to the shop-floor manager for approval — 4 hours (who may be on vacation)
4. The shop-floor manager approves, the form goes to procurement — 2 hours
5. The procurement manager checks the supplier contract — 1 hour
6. The procurement manager calls the supplier and agrees on price and quantity — 1 hour
7. The supplier sends an invoice; the procurement manager clears payment with finance — 2 hours
8. Finance transfers the money — 2 hours
9. The supplier ships the material — 3 days
10. The material arrives at the warehouse — 1 day
11. The material is routed to production — 1 hour

Total: from the moment the material is needed to the moment it reaches production — 4-5 days, of which 13+ working hours

Step 3: Measure the key metrics

For each process, calculate:

• Lead time (cycle time): from start to finish, in working hours and calendar days. Example: procurement — 4 days (13 working hours)
• Cost per unit: headcount × hours × rate. Example: 13 hours × 3 people (production, procurement, finance) × ₽500/hour = ₽19,500. That's the cost of a single purchase
• Quality: how many reworks, errors, and complaints the process generates. Example: 5% of purchases contain errors (wrong quantity, wrong material) and require rework
• Throughput: how many units are handled per day. Example: a procurement manager can process 5 orders a day

Step 4: Identify the bottlenecks

In your map, find the points where the most time and money are lost:

• Shop-floor manager approval (4 hours): a bottleneck. The manager is often unavailable. This is nearly half the entire process
• Finance sign-off (2 hours): a bottleneck. Finance runs on its own schedule, so approval can drag
• Manual form entry (30 minutes): a bottleneck. The form lives in Excel and can be automated

Don't get lost in the small stuff — go after the biggest bottlenecks, because that's where the real upside is hiding.

Stage 2: Root Cause Analysis

The Five Whys

For each bottleneck, ask "why" five times:

Problem: shop-floor manager approval takes 4 hours

1. Why 4 hours? Because the manager is often unavailable or busy
2. Why are they unavailable? Because there's no system to remind them about approvals, and they can't keep everything in their head
3. Why is there no system? Because the company runs on Excel, and Excel doesn't send reminders
4. Why Excel? Because it has always been done this way and no one proposed an alternative
5. Why did no one propose one? Because the process seems unimportant ("it sort of works") and there's no budget for improvements

Root cause: the absence of a system to track approvals + the low priority assigned to this process

Talk to the people in the process

No outside expert knows a process the way the people who live it every day do. So just ask them:

• Production: "How often do you run short on material because procurement is delayed? How badly does it derail the plan?"
• Procurement manager: "Which steps in the procurement process feel inefficient to you? Where do you lose the most time?"
• Shop-floor manager: "What's stopping you from approving purchases faster?"

Often people will say: "We need to automate approvals," or "The manager should approve same-day, not put it off," or "Give the procurement manager authority to decide on anything under ₽100K."

Stage 3: Choose your optimization methods

Method 1: Lean/Kaizen (continuous small improvement)

The idea: every week the team looks for ways to work faster and more efficiently.

• How to implement: weekly 30-minute meetings with the people in the process. Each one proposes a single improvement
• Example improvements: remove a redundant approval step, automate a notification, give the manager a mobile alert for quick sign-off
• Investment: minimal (meetings, perhaps a small spend on tooling)
• Result: typically a 10-20% improvement in 1-2 months, 30% at most
• Time to implement: fast (weeks), with results visible immediately

Applied to our procurement example:

• Week 1: removed the finance sign-off step (gave the procurement manager authority). Lead time dropped by 2 hours
• Week 2: automated the alert to the shop-floor manager (the form now lands in Slack/Telegram). Approval sped up by 2 hours
• Week 3: the shop-floor manager can now approve via Slack (one click). Approval went from 4 hours to 15 minutes
• Result after one month: lead time fell from 4-5 days to 1-2 days (a 60-75% improvement)

Method 2: BPR — Business Process Reengineering (a full rebuild)

The idea: don't improve the existing process — rebuild it from scratch, forgetting the old constraints.

• How to implement: assemble a team and ask, "If we were designing this process from scratch, how would we do it? What systems, people, and procedures would we need?"
• BPR example for procurement:
  • Old process: manual entry, layers of approvals, 4-5 days
  • New process: production plans purchases a month ahead (a priori); the system automatically orders from the supplier when material runs low; approval is required only if the deviation from plan exceeds 20%
  • Result: the process shrinks from 4 days to a few hours, with approvals almost entirely eliminated
• Investment: medium-to-high (a system is needed, plus training)
• Result: a 40-60% improvement, sometimes as much as 80%
• Time to implement: slower (a month or two to roll out)

Method 3: Automation

The idea: automate routine operations (RPA, processing systems, application integration).

• When to use it: AFTER you've optimized the process. Don't automate a broken process — that only makes the problem worse
• Automation examples:
  • Excel form → a CRM that sends notifications automatically
  • Manual invoicing → a system that issues the invoice automatically when an order is created
  • Manual data checks → a system that validates data automatically
• Investment: medium (buying the system + integration)
• Result: a 30-50% reduction in labor on the step
• ROI: typically pays back in 6-12 months

Stage 4: Implementing change (change management)

Phase 1: Preparation (Week 1)

• Pick a pilot group: a small group (5-10 people), the most motivated and open to change
• Explain WHY everything is changing: people won't accept change under the banner of "we have a new process now." They need a reason that lands in human terms: "Right now procurement drags on for 4 days; soon it'll be 1. That means we stop missing plans and firefighting over material shortages."
• Train them on the new process: a meeting, a demo, hands-on practice
• Reinforce the goal: what to measure, and how everyone will know it's working

Phase 2: Pilot launch (Weeks 2-3)

• Run the new process with the pilot group
• Daily 15-minute stand-ups: what's working, where the problems are, what's blocking
• Expect things to slow down at first: in the first week the new process will almost certainly lag the old one — people are still learning. That's completely normal, so don't panic
• Document the problems: what's not working, and why
• Solve problems fast: if there's a system issue, fix it same-day (don't wait a week)

Phase 3: Rollout and scaling (Week 4+)

• Measure the pilot results: lead time, cost, quality. Compare against the targets
• If the results are good: roll out to the whole unit (or if they're poor, return to Phase 1 and rebuild)
• Train everyone else: meetings, documentation, video
• Assign an owner: someone accountable for adherence to the new process
• Document the process: a written guide, video, checklist

Stage 5: Monitoring and improvement (ongoing)

What to measure

• Lead time: from start to finish of the process (should fall by the target amount)
• Cost per unit: the cost of the process per operation (should decrease)
• Quality: errors, reworks (should drop)
• Employee satisfaction: ask people whether the work is easier now (should rise)

How often to review

• Months 1-2: weekly meetings (people adapt, problems surface)
• Months 3-6: weekly meetings (hunt for new gains via Lean)
• Month 6+: once a month (maintenance, monitoring)

Metrics of optimization success

The project counts as a success if you hit at least 3 of 4:

• Cycle time (lead time): down by 20%+ (the usual target is 30-50%)
• Cost (cost per unit): down by 15%+ (the usual target is 20-30%)
• Quality (defects/errors): down by 50%+
• Employee satisfaction: up (survey, observation, absence of complaints)

Common mistakes to avoid

• Mistake 1: Automating the wrong process. You didn't optimize the process — you automated it outright. The result: you've automated 5 unnecessary steps. The right order is: optimize first, then automate
• Mistake 2: No executive sponsorship. The CEO doesn't treat it as a priority, doesn't show up to meetings, doesn't allocate resources. The result: people revert to the old process within a month
• Mistake 3: Ignoring the people in the process. You brought in an outside consultant who elegantly redesigned the process — and not one of the people who actually run it lifted a finger to help build it. No surprise, then, that they simply didn't adopt it
• Mistake 4: Expecting a miracle in a week. The process took 5 years to form; you can't change it in a week. Typically it takes a month to roll out, a month to stabilize, and another month to wring out every last result
• Mistake 5: Picking too large a process to optimize. Start small (a single sub-process), get a result, then scale

Examples of results from our practice

• Manufacturing (500+ employees): optimized materials procurement. Cut lead time from 4 days to 1 day (a 75% improvement). Savings from faster capital turnover: ₽50M+ a year
• E-commerce (200+ employees): optimized order processing. Lead time from 2 days to 4 hours (a 95% improvement). They can now handle 10x the orders with the same team
• Consulting (50 employees): optimized the sales process. The deal cycle shrank from 3 months to 6 weeks (a 30% improvement). Revenue rose 25% with the same team

Conclusion

Business process optimization is systematic work over 90-180 days. Start with the right diagnostics (find the bottlenecks), choose the right method (Lean for small gains, BPR for radical ones), implement it properly (pilot → scale), and measure the results.

Even if you lift just one core process by 30%, the whole company will feel it: costs go down, margins go up, and people can finally breathe easier. So don't take on everything at once. Choose one process. Drive it to a result. Only then move to the next.

One last thing. Good optimization is almost never visible in people moving faster — it's visible in things suddenly getting easier for them: the redundant approvals are gone, requests stop falling through the cracks, the endless "who's responsible for this?" disappears. If the work feels lighter after the change rather than more anxious, you've done it right.

If this sounds like your situation — see how we do this as a service: optimising a company's business processes.