My Story
In my product, we realized we had a very complex process for customer support and incident management.
How did we know?
Because when we asked ourselves "what is the process?" no one could answer clearly.
And because the product is complex, support was handled by four different teams.
You can already imagine what that looks like. Users getting ping-ponged from one team to another. No clear answers. Phrases like "it's not in our scope, try them."
And because we care about user experience in every aspect, we wanted to change that. Build trust. Get fewer escalations. Stop users from going up the chain of command out of frustration.
Luckily, we had some process experts we could work with to untangle this spaghetti.
Here's how we did it.
The Approach (Simplified)
→ Map out your current process (with time consumed + % right the first time)
→ Run a workshop to analyze the flow and create a better one
→ Implement the improvements (the hard part)
→ Measure the results
That's it. Let's go deeper.
Step 1: Map What You Have Now
Start with the current flow. Capture how it really works: beginning to end. No sugarcoating.
Here's a hypothetical example: a user access process for a data product.
The scenario: A new user needs access to a database with sales data.
Look at this. 2.5 to 3.5 hours for a single access request. And only 25–35% of users get it right on their first attempt.
The biggest pain points? Step 3 (user tries to define tables but has no idea what they need) and Step 6 (reading complex data documentation). Those two steps alone eat over an hour and have the lowest success rates.
This is the kind of thing that looks "fine" on paper but is secretly draining your users and your team.
Step 2: Run a Workshop
We did a full-day workshop for this. Here's what actually happens in it.
Workshop Part 1: Surface the pain
Start by letting people talk. What's broken? What frustrates them? What do users complain about?
This isn't just venting. You're collecting signals you probably missed when you mapped the process alone. Someone will say "oh, and there's also this step where..." and suddenly your flow has two more boxes.
Workshop Part 2: Pressure-test every step
Go through each step of your current process and ask:
→ Does this step need to exist?
→ Why does it go wrong so often?
→ Can it be moved earlier, later, or run in parallel with something else?
→ Is there a handoff here that causes delays?
This is where you'll find the real waste. Steps that exist because "we've always done it this way." Steps that duplicate work across teams. Steps that only exist because an earlier step is broken.
Workshop Part 3: Design the ideal flow
If you could start from scratch, how would this process work?
Draw it out. Don't worry about constraints yet. Just design what "good" looks like.
Then compare it to your current flow and list the gaps. What's different? What's missing? What needs to change?
Workshop Part 4: Prioritize the gaps
Take every gap you identified and map it on an effort vs. impact matrix.
→ X-axis: effort (small to large)
→ Y-axis: impact (low to high)
This tells you what to do first, what to plan for later, and what to skip.
Here's a sample agenda you can steal:
A note on timing: this took us a full day. If your process is simpler, you might get through it in half a day. But don't rush it. The conversations that happen between the exercises are often where the best insights come from.
The Ideal Flow
Here's what the improved user access process could look like:
From 3+ hours to 20 minutes. From 30% accuracy to 85%.
That's not a fantasy. That's what happens when you remove unnecessary steps, add templates, and automate the boring parts.
The Gaps
When you compare the current flow to the ideal, you'll spot specific improvements. For our user access example:
→ Create role-based access templates so users don't have to guess what tables they need
→ Build a quick-start guide with common data tables and use cases for new users
→ Pre-fill the PM review with user context so the PM doesn't start from scratch
→ Auto-route tickets based on data sensitivity level instead of manual triage
→ Build self-service documentation with interactive examples, not just raw data dictionaries
→ Auto-provision access after PM approval instead of waiting for manual setup
Now the question becomes: where do you start?
That's where the effort vs. impact matrix comes in.
Start with the quick wins in the top-left. A quick-start guide and pre-filling PM reviews are low effort and high impact. Those alone will save hours of back-and-forth.
Then move to the strategic bets: role-based templates and self-service docs take more effort but fundamentally change the experience.
Step 3: Implementation (The Hard Part)
You've got your ideal flow and your prioritized improvements. Now comes the part where most process improvements die: actually doing them.
Here's what I've learned works:
→ Start with one improvement, not five. Pick the highest-impact, lowest-effort item from your matrix and ship it first. Quick wins build momentum and buy-in.
→ Assign a clear owner. Process improvements without an owner become "we should do that someday." That someday never comes.
→ Set a timeline. Even if it's rough. "We'll have the quick-start guide live in 2 weeks" is better than "we'll get to it."
→ Communicate the change. Tell your users and your team what's changing and why. A new process nobody knows about is the same as no process at all.
→ Expect resistance. People are used to the old way, even if it's broken. Be patient. Be persistent.
Step 4: Measure
How do you know if it actually worked?
Go back to the same metrics you captured in Step 1:
→ Time consumed: Is the process faster? Track total time from start to finish.
→ % right the first time: Are users getting it right more often? Fewer rejections = less rework for everyone.
→ Number of escalations: Are users still going up the chain, or are they getting answers within the process?
→ User satisfaction: A simple "how was your experience?" after the process. Even a thumbs up / thumbs down works.
→ Volume of support tickets: If you fixed the right thing, this number should go down.
Compare your before and after. If the numbers improved: celebrate, document, and move to the next improvement. If they didn't: go back and figure out why.
The goal isn't perfection. It's progress.
