Frustration Friday.

For the unsung heroes and teams iterating through ambiguity, rework, and shifting priorities.

We see you.

Now let’s go ship one thing that matters.

Then product velocity stalled.
The business could no longer justify the burn.

Too many leaders recognize this pattern too late.

Is your team in the Doom Loop?

Too often I’ve seen CEOs preside over major rewrites which slid from hope to frustration and alarm. Whole teams got laid off as a result.

Many of you reading this have lived it. Perhaps more than once. You doubled down, while others bailed.

I’ve seen startups and middle market companies alike freeze a working product to chase a rewrite.

Velocity drops. Parity drags. Burn spikes. The team is capable. But it's running two systems and shipping neither.

Most rewrites don’t fail suddenly. They drain momentum over time.

#shipgoodproduct

Strong product leadership prevents software rewrite pressure.

The key signals of healthy modernization: 1) clear priorities, 2) aligned teams, and 3) continuous customer value.

These disciplines protect organizations from the temptation of risky resets.

#mission
#priorities

Software rewrites succeed when leadership understands the storm they’re entering.

Velocity loss is often dismissed as temporary. It isn’t.

If you can’t keep improving customer value during the rewrite, growth stalls.

Some companies weather it. Others fragment and sink.

#rewritesareexistential

When orgs stop evolving their software products incrementally, their Rewrite Gravity Index™ grows.

To find out your risk, do a quick appraisal at:
appraise.mandau.co

It's like a 20-point safety check for your vehicle, but for something much more precious. Your business.

100%! What stops teams from doing this?

@nonzerosumjames.bsky.social ty

@effectivealtruist.bsky.social ty

The danger isn’t the rewrite itself.
It’s how many risks go live at the same time.

Architecture shifts. Data migrations get challenged. Customer transitions stall.

Navigating the existential danger means knowing how to reduce unnecessary, compounding risk.

#unnecessaryrisk

LLMs feel intelligent, but they still struggle to stay on task.

They reconstruct knowledge, not retrieve it.

Strong AI products are built by focusing context, whether by improving models or improving their usage in the product loop.

#attention
#productsense

Incremental progress compounds gains while dissipating risk.

Shipping small improvements regularly creates two advantages:
1) customer trust grows, and
2) risk stays contained.

Organizations that preserve momentum rarely need sweeping rewrites.

#customerfirst
#peakproductdiscipline

Yes, this! Compound wins are like compound interest! Compound risks work the same way, ... but in reverse.

That's a great insight, @jeremiahchronister.bsky.social and makes sense! The broader pattern holds in other areas like changing to completely different industries or even sports, unless common threads are present.

The most expensive rewrite freezes progress.

When teams stop shipping while waiting for a big-bang release, they lose the compounding value of continuous delivery.

Standish Group found a 20-fold increase in success rates when teams ship in small increments.

Momentum matters.

#continuousdelivery

Sometimes a software rewrite becomes the only remaining option.

If that existential moment arrives, success hinges on one thing: continue to give your customers outstanding product value.

The goal is not to start over.

The goal is to preserve momentum during the change.

#nocodefreeze

The strongest software organizations continue to evolve.

“We change our [search] algorithm almost every day.”
– Amit Singhal, Google Search (2011)

Small improvements compound.
Disciplined continuous evolution prevents product debt & technical debt from reaching existential levels.

#modernization

The character limit is really restricting to make this nuanced point, but I have a set of slides on my website (link in bio), which makes the point better. Also an executive white paper on rewrites with documented stats and quotes, for anyone interested. Only 9 pages, but it's a page-turner :).

We need to raise awareness!

To be fair, rewrites aren't always the worst option, just usually. For example, Facebook rewrote their frontend a decade ago, twice. The 1st one was existentially bad. The 2nd one undid the damage of the 1st but to fend off disruption, FB had to acquire Instagram.

Real question, how would you share this information?

Relevant case studies on the specific topic of software rewrites, which the case studies address. Spam is irrelevant or excessive. E.g. 20 links to cat videos in the same post on a different topic. Each original reply was on a completely different thread. I'll take your word on the counts.

:) I see them. What I mean is even obvious bugs QA may flag may not be enough to stop a release if the value of releasing is greater than not releasing. An easy example is if a bigger bug is addressed in the release. I didn't see this context in your post, so I saw an opportunity to contribute.

All the posts are discussing software rewrites. And I tailored a lot of those replies to the specific conversation. I'm not posting cat videos on random threads here. Maybe I caught you on a bad day.

Thanks for your message, Eric. The case studies are relevant, but I can see how you could interpret it that way. Curious, wouldn't you want the conversation to continue? It's a growing and timeless problem.

Takes time to catch up on sleep (product debt). Can't do it all in one night, and can't sleep for a year straight either (big bang). But when companies try to both sleep and walk (code freeze and rewrite legacy), that's the fastest way to failure on both fronts.
mandau.co/case_studies...

Continuous is the way, but to get there from a bad place is tricky. Fortunately, help is available.
mandau.co

Usually a bad idea, but it's a big grab bag of activities so context matters.
mandau.co

mandau.co/case_studies...

Need context, but yes, this could be true.
mandau.co

100% !

There's often a much less expensive, safer, and more effective way which won't tank your company.
mandau.co