You've seen this project. Everyone nods along. "Yeah, we know what we're building." The roadmap looks clean. The timeline feels reasonable. Vibes are good.
Then someone asks one clarifying question and the whole thing unravels. Turns out nobody agreed on anything. You just used the same words differently.
By the time you realize the mess, it's too late. Half the budget is gone. Management is asking hard questions. Your team is demoralized and finger-pointing starts.
Stop assuming alignment. Prove it instead.
Projects aren't complicated. They're follow-throughs. The hard part is knowing where you are and where you're going. Everything else is just the diff between the two.
What is AB-Diff?
Define Point A (where you are). Define Point B (where you're going). Calculate the difference. That difference is your backlog.
Simple in theory. Brutal in practice. The anxiety is the point.
Step 1: Define Point A
This is the brutally honest inventory. No optimism allowed.
You need to document what actually exists today. Not what should exist. Not what you thought existed. What's really there. Who is actually involved and what do they actually own? What decisions have been made versus assumed? What technical debt lives here that everyone pretends doesn't?
The trap is brushing off ambiguities as obvious. "Oh everyone knows what that means." No they don't. When someone says "you know what I mean" — stop. You don't. Make them spell it out.
Every ambiguity is a landmine waiting to blow up your timeline three months from now. Treat them with anxiety. Write them down. Resolve them before moving forward.
Get everyone in a room. Force the uncomfortable conversations. The goal isn't consensus on what to do. It's consensus on what is.
Step 2: Define Point B
This is not what you're building. This is what you're achieving and why.
Focus on the outcome, not the implementation. What does success look like in (preferably measurable) concrete terms? Why does this matter to the business?
Avoid vague destinations like "Make X better" (better how?), "Modernize Y" (to what standard?), or "Build v2" (v2 of what exactly?). These accumulate scope forever and won't let you move on. Instead focus on clear business outcomes like "Migrate A service to B database by C date to achieve D cost reduction" or "Implement A feature by B date to test if it increases user engagement". Deciding what not to do is more important than deciding what to do.
The trap here is jumping to HOW before you've nailed WHY. Your team figures out the how. Your job is the why. The why tells you if Point B is even worth reaching. It guides every decision downstream.
Apply the same anxiety you gave Point A. If you can't explain Point B successfully to a junior for any reason, you don't have Point B yet. You have wishful thinking and a migraine scheduled a few months from now.
Step 3: Calculate the Diff
Now work backwards from B to A. What has to change to get there? That list is your backlog.
This diff should be exhaustive. Nothing left implied. Nothing assumed. It should contain everything the layer below you needs to complete these tasks without coming back to ask what you meant.
Break it into pieces your team can execute. Make ownership clear for every item. Make outcomes testable so you know when each piece is done. Your job as a leader is to think through the hard, ambiguous, cross-cutting concerns so your team doesn't have to. Give them clear, bounded work.
Their job is execution. You've already done the strategic thinking.
Step 4: Track and Validate
Project tracking becomes almost trivial now. You have subtasks. Either they're done or they're not. No status theater is required.
Here's the validation test: If completing every item on this list does not get you from Point A to Point B, one of three things went wrong.
Either you defined Point A incorrectly, you defined Point B incorrectly, or your diff wasn't thorough enough.
When new information arrives (oh, and it will) update the model. Re-diff. This isn't failure. This is the process working as intended. You're catching problems while you can still adapt, not after management is already upset.
Take the moment to ask yourself and your team why those items were missed. Were they legitimate blind spots only shipping could find? Were they bad assumptions you held? What do you need to put in place so your next AB-Diff is even better.
Why This Works
This forces alignment before a single line of code. No more "I thought we agreed" three months in.
This also surfaces hidden work early. All those systems nobody remembered, integrations everyone assumed would just work, ownership gaps between teams show up in the diff, not in a production incident.
Naturally, this makes ownership explicit too. Every item has a name next to it. No more "I thought you were handling that."
Management couldn't ask for clearer visibility. They can see the backlog. They can see progress. No weekly status meetings where everyone performs confidence.
Scope creep is now super obvious. When someone asks for something new, you have an answer: "That's not in the diff. Want to re-scope Point B?"
Story Time
A client hired my Byte Bot a legacy migration. They were confident they knew what to do. They even had a timeline.
Despite hesitation, we made them define Point A and Point B. Just that exercise, before any planning, revealed a ton of work nobody had considered. They even forgot an entire system existed. Deployments they assumed would "just work."
Then we calculated the diff. Their original estimates would have been wildly off. Not just by a little off. We are talking months off.
We caught this super early, months before management would have started asking hard questions. The client adapted their timeline, adjusted their budget, and set realistic expectations with stakeholders.
The project still had challenges, but they were challenges we chose, not surprises we stumbled into. The team stayed sane. Management had visibility without micromanaging. When the board asked about progress, there was a clear answer with numbers.
Needless to say, they extended the contract with us. They saw as too valuable.
Closing
You can't plan your way out of uncertainty but you can shrink it dramatically by forcing clarity upfront.
You define where you are. You define where you're going. You calculate the gap.
Everything else is follow-through.
