Why Starting Over Is Sometimes the Fastest Path Forward
Editorial note: Originally published in October 2012. This version has been extensively rewritten and updated in 2025 to align with how I now think about rework, learning, and building effective systems.
“What do you mean you’re throwing the prototype away?”
Wally shook his head, “Not throwing it away, it did its job. Now it’s time to make it work at scale.”
“Are you crazy?” I squeaked. “We’re a month behind schedule, and you’re going to start over?” I pointed at the prototype, sitting on my desk. “You’ve got a working model right there. You take it back to your mad scientist lab, and you make that work. Sales has a big demo next week.”
I watched Wally dejectedly grab the prototype and head for his lab. “Seriously,” I said to the ceiling, “what part of move fast doesn’t anyone understand? Rework the whole hardware design? What next, we’re going to rearchitect the database.”
“Probably wouldn’t be a bad idea,” a deep voice said. “It was built decades ago for a product that didn’t have half the technology we do today.”
“Go away, Hogarth,” I said without looking down from the ceiling tiles. Staring at acoustical tiles can be so relaxing, especially when the alternative is to talk to the gorilla in the room.
I could hear him shuffling across the floor (do gorillas generate static electricity when they shuffle?). Hogarth said, “Seriously, Wally has the right idea.”
I dropped my gaze to look at Hogarth. “It doesn’t matter, we’re too far behind schedule. We’ve got something that works already, no need to reinvent the wheel.”
Hogarth shrugged as he dropped into a chair. “If you think about it, we’ve reinvented the wheel a lot. The first ones were made of stone.”
“Why are you here anyway?” I snapped. “If you’re here to pester me about the prototype, forget it. It works, and we just need to tweak it to make it ready to ship.”
“Oh, right,” Hogarth said. He laid a sheaf of papers on my desk. I don’t know where he produced them from, and I’m pretty sure I didn’t want to know. “I’ve been writing a book, and I wanted you to look at it.” He held up a massive hand, “Sorry, it’s in longhand. Have you ever tried to use a human keyboard with fingers this big?”
“A book?”
Hogarth nodded, “Yep, ‘Conversations With The Gorilla In the Room’.”
I took the stack of papers from him and began looking them over. A riot of text instantly assaulted my eyes. The handwriting was criss-crossed with struck-out words, corrections, and even a whole paragraph written sideways in the margins.
Looking up from the papers, I looked at Hogarth. His face was eagerly looking at me for a reaction (Note- Gorilla’s eager looks a lot like “I’m going to eat you”). “Hogarth, this is a mess. I can’t read it, much less follow the narrative.”
Hogarth looked at the paper, “Yeah, well, I really didn’t want to write it all over again. You can still read it, right?”
I boggled at him for a moment. “It’s like you took duct tape and baling wire to your writing. It’s haphazard and inconsistent. I’ve got to turn the page sideways to read half this stuff.”
“Rewriting it would have taken too much time, though,” protested Hogarth.
“Hogarth,” I slipped into my professor voice. “Rewriting is part of the process. Not only does it make your work better, it also makes it easier for others to follow it.”
Hogarth’s eye lit up, “Oh, you mean it will work better? Kind of like the carpentry maxim of ‘Measure twice, cut once’?”
All but standing up in my chair, I replied, “Yes, precisely!…” And then I stopped dead. Looking Hogarth straight in the eyes, I said, “You did it to me again, didn’t you?”
Hogarth pulled a fichus branch from somewhere and began to nibble at it merrily. “Uh-huh, I did.”
Hoisted on my own gorilla, again.
REWORK, REFACTOR, REDO: GO FAST TO GO EVEN FASTER
“Go slow to go fast” is a phrase I’ve heard for most of my career.
It’s usually offered as a warning against recklessness, a reminder that careful planning prevents costly mistakes. There’s truth in that. And it’s also incomplete.
The real question isn’t whether we plan or move fast.
The question is how quickly we’re willing to learn.
Most modern work, especially product development, happens in environments where certainty is low and change is constant. In those conditions, the fastest teams aren’t the ones who avoid rework. They’re the ones who design for it.
They build with the expectation that what they create will teach them something. Sometimes that learning leads to refinement. Sometimes it leads to replacement. And sometimes it reveals that starting over is the most responsible option available.
That’s the part that makes people uncomfortable.
This is a story about intentional rework, and why starting over is sometimes the most disciplined move you can make.
Rework often gets labeled as waste because it looks inefficient. You had something that worked, and now you’re touching it again. From the outside, that feels like backtracking. From inside the system, it’s often the moment where hidden constraints finally surface.
The mistake isn’t rework.
The mistake is pretending early decisions were perfect.
This is where planning is often misunderstood. Planning isn’t a phase you complete so you can finally begin “real work.” Planning is how teams make sense of what they know right now, align on intent, and decide what to try next. As soon as reality provides new information, good plans evolve.
Planning is communication.
Planning is alignment.
Planning is ongoing.
Only planning without building is speculation. And building without revisiting the plan is just momentum masquerading as progress.
That’s why rewriting matters. That’s why refactoring matters. That’s why sometimes the right move is to step back, acknowledge what the first version taught you, and build again with clearer eyes.
Carpenters say, “Measure twice, cut once.”
In modern product work, I’d update that slightly: