The Most Common Mistake is to Optimize a Thing That Should Not Exist

After attending the White Out game against Oregon, I've been thinking about this a lot lately.  James Franklin will not get Penn State over the hump.  He’s good for 10 wins and a Top 10 finish every season, but will always lose to other Top 10 teams not named SMU or Indiana (which, if we’re being honest, don’t really count).

Before he went off the rails, Elon Musk said, "The most common mistake of a smart engineer is to optimize a thing that should not exist."  Penn State is optimizing a thing that should not exist - James Franklin as Head Coach.  Instead, they should be focused on building a true contender.  His ceiling is below the floor of Smart/Lanning and even Dabo.

I see the same pattern in commercial real estate. We're really, really good at making broken processes slightly less broken instead of building the right thing from scratch.

We improve our spreadsheets. We create more detailed checklists. We hire coordinators to coordinate the coordinators. We have quarterly meetings to discuss how to make next quarter's meetings more efficient.

We optimize things that shouldn't exist in the first place.

For example, a few years ago we built these great Excel spreadsheets to help clients calculate NOI adjusted DSCR for covenant tracking.  We already had very complex cashflow models. We could quickly drop in forward curves every quarter - we just needed the T12s and balance sheets.  Toss in some customizable input fields to make the NOI adjustments and we had a pretty efficient covenant tracking spreadsheet. 

But the ceiling for those Excel spreadsheets was below the floor of the right solution.   We were optimizing the wrong thing instead of building the right thing.

So we built it as software.

It was brutal.  Every single possible adjustment we could find across 5,000+ loans.  Sure, adjusting for management fees or vacancy comparisons is easy enough, but what about “greater of” T12 vs T3 income and expense comparisons?  Hypothetical amortizations with Treasury lookbacks, rent roll exclusions…it took years and we still work on improving it every day.

Fully customizable, fully automated, 100% accurate down to the penny.  And we are the only company in the world that can do it.  

Excel was never going to get there.  We had to trust that it was better to build the right thing than to keep optimizing the wrong thing.  I had to bite the bullet and commit, trading out short term pain for long term gain.

AI makes a tough decision even tougher.  What if building something the right way today will be rendered worthless a year from now when AI says, “Hey, I can do that for you for $20 a month!”  I don’t have a good answer for that.

The Real Boss™ likes to say, “Good is the enemy of great.”  AI might muddy the water, but doesn't change the underlying philosophy.  

The most common mistake is to optimize a thing that should not exist.

And if you’re wondering, “Why isn’t JP mentioning that embarrassing loss to UCLA yesterday?”, that’s because I wrote this last Sunday and forgot to hit send.  In the meantime, PSU kept optimizing a thing that should not exist and only reinforces my feelings.