How to build a process improvement plan

A process improvement plan spells out what’s broken in your current workflow and how you’re going to fix it. That’s it. No mystery. Here’s how we approach process improvement at Tallyfy.

What a process improvement plan really is

Here’s where most articles overcomplicate things. A process is just a sequence of steps and decisions that get a specific piece of work done. Onboarding a new hire. Closing out payroll. Approving a purchase order. Some processes are so routine you don’t even notice them. Others are so critical that if they stopped, the entire operation would grind to a halt.

A process improvement plan documents your strategy for making one of those sequences better. Not reactively, like scrambling when something catches fire. Proactively - looking at your workflows, finding where things get stuck, and fixing them before they become emergencies.

Almost all quality improvement comes via simplification of design, manufacturing… layout, processes, and procedures.

Tom Peters (Source)

The plan itself typically answers questions like: which process did we pick and why? How are we measuring it? Who’s on the team? What resources do we need? How do we roll out the changes and make sure they stick?

That last question is where things sort of go sideways for most organizations.

The question we get asked most often with workflow automation, the planning part isn’t what trips people up. It’s the sticking part. I’ve seen teams build beautiful improvement plans with perfect data and flawless testing - only to watch adoption crater within weeks of launch.

Ten-step approach that works

There are plenty of approaches out there. DMAIC from Six Sigma. Lean. Masaaki Imai’s Kaizen. They all have merit. But at their core, they follow a similar arc. Here’s a ten-step version that pulls from the best of them.

Step 1 - Pick the process and define what “better” looks like. Be specific. “Improve onboarding” is too vague. “Reduce time-to-productivity for new hires from 90 days to 45 days” gives you something to measure.

Step 2 - Assemble a small, focused team. Not a committee. Include people who actually do the work, not just managers who think they understand it.

Step 3 - Map the current process visually. Every step. Every handoff. Every decision point. You can’t improve what you can’t see. This is where Tallyfy helps teams skip the flowchart phase entirely - you map the process by building it, and it becomes the living workflow people follow.

Step 4 - Collect baseline data. How long does each step take? Where do things pile up? What’s the error rate? Without this, you’re guessing. Align the data with your workflow so you can see what’s happening in context.

Step 5 - Figure out if the process is even stable. Wild variation tells you something fundamental is wrong. You might need to stabilize before you can improve.

Step 6 - Compare your data against the improvement objective. Can the process, as designed, ever hit the target? This is where you uncover the real bottlenecks.

Step 7 - Identify the primary blocker. Not all problems are equal. Which single issue, if solved, would have the biggest impact on reaching your objective?

Step 8 - Build a specific change plan. What changes, who owns each change, and by when.

Step 9 - Test the new process. Collect fresh data. Compare against the target. Iterate.

Step 10 - Assess whether the changed process is stable, capable, and genuinely improved. Look at operating costs, labor time, error rates, and throughput.

This isn’t a one-and-done exercise. It’s a cycle. Which brings me to why most plans die.

Example Procedure

Quarterly Strategic Planning & Goal Setting Workflow

1Revisit annual plan goals

2Break down goals into smaller chunks

3Review budget and benchmarks

4Create action steps and benchmarks

5Set expectations and timelines

+2 more steps

View template

Example Procedure

Annual Planning

1Define your goals using SMART criteria

2Build your budget and financial projections

3Set timelines and quarterly checkpoints

4Create contingency plans for when things go wrong

5Review and finalize the annual plan

View template

Why most process improvement plans fail

You might expect the answer to be bad data or a weak method. Turns out, it’s not. Research shows that roughly 70% of process improvement initiatives fail, and the primary reason is people.

The answer: People.

OK, that’s a bit reductive. This always surprises teams. Processes don’t deviate on their own. Since they can’t change themselves, that means people change them - or resist the changes entirely. Someone decides to skip a step because they’re busy. A manager overrides a rule because “it’s just this once.” A new hire never learns the original process and invents their own version. Within six months, you’ve got five messy versions of the same workflow running in parallel and nobody realizes it until something breaks badly enough to force a review.

From what we’ve heard in hundreds of conversations across mid-market and enterprise organizations, the same patterns keep surfacing:

• Communication disconnects - The team building the plan and the team executing the work aren’t talking enough
• Executive expectations vs. reality - Leadership expects overnight transformation while the people doing the work know it takes months
• No input from frontline workers - Plans built in conference rooms by people who haven’t touched the actual process in years
• Ripple effects ignored - Improving one process disrupts three others nobody thought about
• Budget gets pulled - Leadership gets excited about the next initiative before this one is embedded
• No feedback loops - Nothing gets measured after launch, so nobody knows if it’s working

A consulting firm COO once told me they evaluated “dozens of applications” before finding the right combination of functionality and ease-of-use. Their core problem? Multiple people needed to access and track the same processes to make sure steps happened in order. Without that visibility, things got missed.

If you want a process improvement plan to survive contact with reality, you need to focus on the people more than the process itself.

Two approaches that make improvements stick

Visual thinking

When you want people to change how they work, show them. Don’t just send a PDF.

This matters more than most leaders realize. The Picture Superiority Effect, backed by Allan Paivio’s dual-coding theory, demonstrates that people remember roughly 65% of visual information after three days, compared to about 10% of what they only hear. Images get encoded through both visual and verbal pathways in the brain, creating a stronger memory trace. Kind of a big deal, that gap.

For process improvement, this means your teams can see exactly what they’re supposed to do and how their piece connects to the bigger picture. It makes people feel valued, too - understanding how their work impacts the entire chain builds engagement.

This is part of why we built Tallyfy the way we did. Instead of static flowcharts nobody looks at twice, Tallyfy turns the process into something people interact with daily. The process isn’t a document sitting in a shared drive. It’s the actual system guiding the work.

Co-creation

I’m not convinced that hiring expensive consultants to redesign your processes in a conference room produces lasting results. In our experience, the people doing the work every day know exactly where the biggest problems are. Co-creation means involving those people in designing the new process. Not just collecting their feedback and ignoring it - genuinely incorporating their ideas. One IT consulting firm built their entire onboarding process with checkpoints at 30, 60, and 90 days, plus a six-month review. They didn’t hire an outside firm to design it. The HR manager who ran onboarding every week mapped it out - nineteen steps before day one, then sixteen more through the first year. Research on participatory design confirms what we’ve seen at Tallyfy: designers create better solutions when working alongside the people who’ll use them. And since people tend to support what they helped build, adoption rates go up significantly.

The combination of visual process design and co-creation is powerful. Give teams a tool where they can build and modify the process together, see it visually, and run it in real time - that’s when improvements actually stick.

AI and process improvement - a hard truth

Here’s the mega trend I keep coming back to: AI doesn’t redesign your process. It runs it at 10x speed, flaws included.

research that over 40% of agentic AI projects will be canceled by 2027 because of escalating costs and unclear business value. McKinsey’s research confirms that organizations reporting significant financial returns from AI are twice as likely to have redesigned end-to-end workflows before selecting their AI tools.

Think about that. The organizations winning with AI aren’t the ones with the fanciest models. They’re the ones who fixed their processes first.

If your approval workflow has six unnecessary steps and three redundant handoffs, automating it with AI just means you’ll execute a broken process faster. You’ll get the wrong answer at scale, with impressive speed and consistency.

After watching hundreds of teams try this this play out repeatedly. Teams come to us saying “we want to automate our processes with AI.” The first question we ask is: “Have you mapped and improved the process yet?” Usually the answer is no. That’s where the real work starts.

Process definition and standardization is the prerequisite for any meaningful AI adoption. Get the process right. Then automate. Not the other way around.

Continuous improvement after the plan

Once you’ve rolled out your process improvement plan, the work doesn’t stop. Is there a finish line? No. Take a position of constant improvement. Regularly audit the new process and gather feedback from your teams.

The best organizations treat process improvement as an ongoing discipline, not a project with a start and end date. Small, continuous tweaks beat massive overhauls every time. That’s the core insight behind Kaizen, and it holds up.

Use the data. Watch for drift. Keep the feedback loops alive. That’s how you move from improving processes to building an organization that improves itself.

Common questions about process improvement plans

What should a process improvement plan include?

Clear goals, an analysis of the current process, identified areas for improvement, action steps with owners and deadlines, measurable targets, and a timeline. Think of it as a roadmap - each piece is specific and connects to the next. Without measurable targets, you’re just writing rubbish on paper.

What are the five stages of DMAIC?

The DMAIC model stands for Define, Measure, Analyze, Improve, and Control. It’s the most widely used Six Sigma approach. You set your goals, measure where you are today, analyze what needs changing, make improvements, then control the new process to maintain gains. It works for any industry.

What are common types of process improvement plans?

Popular approaches include Six Sigma (reducing variation), Taiichi Ohno’s Lean Manufacturing (removing waste), Total Quality Management or TQM (quality across the organization), and Kaizen (small changes, constantly). They all aim at the same target through different methods. Pick the one that fits your culture - the best approach is the one your team will follow.

Why do process improvement plans fail?

People, not processes. Research shows the top failure factors include resistance to cultural change, weak management support, poor communication, and insufficient training. Plans built without frontline input fail at dramatically higher rates than those co-created with the people doing the work.

Is planning enough without the right tools?

Honestly, no. A plan on paper is just a plan. You need a way to make the improved process the default way work gets done. That means tracking, visibility, and accountability built into the workflow itself - not sitting in a document that collects dust. This is where process improvement software becomes the bridge between planning and execution.