Maintenance as innovation

A parable on how all innovation depends on maintaining what we have

[Part of an ongoing series - one, two, three, four, five, six]

I

Once upon a time, there lived a King who loved to build giant monuments in his honour. In itself, this isn't particularly unusual of course. There are monuments everywhere. A monument to the King's glory, one that would take 60,000 man hours to build. So either you need to stack up 60,000 men in an easily configurable fashion to make this come about in an hour (unlikely unless one's omniscient!), or do some combination of X men and Y hours, where X*Y is 60k man-hours, with their exact arrangement being subject to the coordination costs that exists in the real world.

Another option is to take generations to build, and pass down the unfinished work to your children. Petra reportedly took 450 years, half a millennium, to be built. You better hope, if you're taking this route, that your children and your children's children want to do this task you've set for them. Perhaps enforced by an external authority whose rule isn't subject to sudden and violent change.

Another option is to learn a little bit, and compound that knowledge over time. Say someone realises "oh, every time we put a column up we have to move this giant stone from that end to this end, and tilt it just so. If we built a roller to make the pillars move around faster, we could save 10 man hours." Assuming there are, say, 100 pillars, that realisation is worth 1000 man-hours saved.

But hold on! You still need to build the roller. That takes some time, say 100 man hours all in all. What about ensuring the thing is used properly, add another 100 man hours to train everyone. It needs to be kept in good nick, cleaned and stored and repaired when necessary. Another 100. To run this entire process from start to finish, another 100.

All of a sudden, the 1000 man hours saved has come down to 600. Still a pretty good improvement, all things considered. The smart masons pat themselves, and each other, on the back.

"But if we change the wheels from stone to wood, it will require less maintenance, and can be built faster, and repaired easier. It will save at least 20 man hours." Follow the exact same logic, and we get to another incremental improvement in our ability to build the monument.

Someone goes and tells the King about all the progress. And he goes, this is fantastic. These few weirdos in a separate corner seem to be coming up with all sorts of ways to make my monument get built faster. They've been spending their evenings and burning expensive lamp oil making it happen, maybe I should put them all in a room and ask them to do just this thing.

And boom! "What if instead of moving blocks around through carts, we created a contraption with pulleys and levers that moved a bunch of them at the same time." The team grows in size as they try all sorts of new ideas to find ways to work faster. Some genius looks at a water well and realises water is just another object, albeit a little pliable, so maybe it has lateral applicability.

"What if we built a bunch of those machines together, so that each individual machine only takes a fraction of the time to be made." Building each machine seems to be making the artisans tired and kind of unproductive. But Steve is great at hub and spokes, and Ian is fantastic at pulley systems. George is fantastic at the assembly. Why not let each do what they're good at, and build a bunch of machines all at once.

If you only wanted to make a few monuments, this is kind of perfect. You can get it all done in a short amount of time, and spend the rest in rest and whatever. Like the old hunter gatherers, once all work is done in a short time, you get to spend most of the time just chilling.

But the King didn't get a divine mandate saying "make only four monuments, that's all you need to do." Once four are done, even if his father could only make two, he wants to make even more. There's no ceiling to monument building greatness.

So off you go again. The reward for good work is more work. But it's fine. Everyone gets paid some amount generally commensurate with their contribution. If you're big and strong and healthy and can move post sections twice as fast as others, you get some recognition for that. But those who did things 10x or 100x as better? Well they can't get paid 100x, because the King would run out of gold and have to wage even more wars, but they get a few slivers more for sure. So it's fine. You work again, but it's good. Because now you can employ a bunch of people to do the stuff you never wanted to, like clean your house or cook food or buy robes.

"The biggest problem is that even after building all these things we end up spending hours teaching the yokels to use the damn things. What if we pre-trained the slaves in charge of moving them around, so they already kind of know what needs to be done, and we don't need to spend so much time re-teaching them." So the only choice then is to set up some kind of a system to get the future workers to get better at their jobs. Maybe don't teach them everything, but at least the basics. Don't tread on each others feet. Pulleys are pulled clockwise, not counterclockwise. This is how you winch. This is how you create routes to move within a construction zone. And so on.

"Could my rival King from across the river and the desert send someone and copy all of this?" The King wonders. Unfortunately the answer is yes. Of course you can make the school secret and the digs secret and the machines all covered up, but someone half smart could see through all that. And the King Across The Desert is no dummy. The only solution is to gain a commanding lead. So the schools continue too.

Before you know it, the 60k man-hour building has come down to 10k, and we can build it so much faster. The King is very happy. The workers get extra beer and rest. The overseers get beer and bread and rest. The ones who discovered the pulley system get better houses, and some even mansions, like that team who figured out how to make a spinning wheel make things move in a straight line. I mean, the King doesn't have that many mansions to spare, but these guys seem worth it. Yeah they're a bit of a weird bunch, and seems to be growing oddly by the minute, but still, they get results.

It also ends up in a static place soon enough such that the next level of incremental improvement can only be done by applying an equivalent amount of manpower elsewhere. We are at peak efficiency, and unless there is some kind of a step change (let's say ... anti-gravity) we won't be able to move much further ahead.

But now there is an entire apparatus that has been created, who are exceptional at finding ways to reduce the time to build giant monuments. They also need to be occupied, work created for them, and the only way to ensure it is that giant monuments continue to get built. Let this fly, and soon you end up with a land that's covered with pyramids.

What happens to all those stone carriers and masons who were previously employed by the King and the King's forefathers? They are now surplus, at least for any one particular monument. Unless there's demand for multiple monuments, which they can now do, or maybe more stone quarries (assuming those who carry stones well can also dig them up), they're going to be bang out of luck.

But those who are in charge of figuring out how to save man-hours, they do pretty well for themselves. Come up with an idea or two, and the impact is way beyond what they could do with their muscles or their chisel. Occasionally they would invent something that would save 100s of man hours, and it's only fair to pay someone 10 or 20 man-hour worth when they do that. And if you oversee a team of ten, each of whom can potentially create such efficiencies, it's only fair that you get paid more, maybe 50 man-hours worth. After all, if you do your job well, it's easily conceivable that you'd save many multiples of that for the King.

The stone carriers can get upset, and moan with their fellow masons, but what can they do. It's either chip stone, or get fed to the crocodiles. They're the ones who do the work, but they're also the ones who are completely constrained in what they can contribute! Individual muscles and chisels have an upper bound in what they can create. But if enough of them go on to join the school and become thinkers, maybe their lot would improve.

And so the kingdom endures. The masons chip away during the day and drink homemade beer at night. The overseers and the inventors and the inventor overseers drink mead and honey flavoured beer and sleep on silken sheets. And the crocodiles stay fed.

II

I moved from India to Singapore when I was 17, and I lived there for almost a decade. As an immigrant first though one thing that struck me when I moved to the city state is how new everything looked. As if it had only been built yesterday. It wasn't, of course. At the time I understood this as the difference that manifests itself when the per capita income is different by an order of magnitude plus.

But after a decade I moved to the UK. And as an emigrant, I found something else. Standards of building clearly didn't show correlations on economic terms. The quality which should've been equivalent, was clearly not. Apart from small pockets of similarity, London, with it's vast wealth and culture and know-how, looks grimy and unkempt and in desperate need of a wash, a double whammy considering the proverbial rain.

So it's not just the riches that potentially exists to be used for the purpose, but the actual distribution. So far, so obvious. Not to me, at the time, but at least in retrospect.

The difference lay in the chasm between the activities of building and maintenance. Despite being the peanut butter and chocolate of the property industry, and seemingly inextricably linked in our lexicon, they are opposing forces, even in a Hayekian lens. Those who build are motivated by the vision of making a mark, of completing a project, of seeing reality moulded to their fantasy. Those who maintain feel little of the glory. Theirs is not to construct, to disrupt the status quo, but rather to ensure the continuation of what's already been done.

Every time there's a small blotch on the outside of a building, it gets identified and patched and cleaned. Every time there's rain damage, cosmetic though it may be, buildings get repaired. Every time there's a pothole in the road, it gets treated so thoroughly it becomes hard to figure out if something was even done.

When there's a relatively confined amount of space and buildings, maintenance is something that gets enough attention and visibility, and therefore gets done. When there is just way too much, like in a larger country like the UK, the sheer amount of things that does require maintenance means the assessment and execution of the maintenance tasks itself becomes a massive undertaking. And like all massive undertakings, it too requires regular upkeep. So the cycle continues through enough inefficiencies that the amount that actually gets done becomes minimal.

All of this has a practical use of course. Because depreciation is real, as Charlie Munger is fond of saying. "The idea that depreciation doesn’t matter? … You have to replace the trucks all the time. It’s a real cost. As Warren says, it’s the worst cost, because you pay it way in advance, and get the benefit way later..." So if you don't fix things upfront, the payments due later as up. The interstate highway system cost c.$500 Billion to construct, and the cost to maintain is but a fraction. But it's like insurance, if you don't pay it, the eventual cost only balloons. And just like insurance, the pinch of having to pay now for "no reason" looked large in our consciousness. But unlike insurance, this is an actual physical intertemporal exchange. The value of goods in real life changes dramatically.

Sadly no leader gets elected or selected on maintenance. "I'll keep the lights on" doesn't make a resounding platform, doesn't loosen donor purse strings, doesn't make people want to follow en masse. Only on building. Even the pork inherent in bills tend towards the latter, since it's more visible and easiest to demonstrate. Compare with "I'll electrify the whole state". You can almost feel the future utopia beckoning. But once you've electrified everything, you do need to keep the lights on.

The alternative is the daemon of depreciation at your door, asking for his due.

But it's also a question on where innovation actually takes place. Is a society that hoards its capital and uses it for maintenance going to be better positioned than one that's not? Presumably there answer lies in the ability of that marginal dollar to create excess wealth in the time between the spend and the depreciation daemon collection date.

There used to be a shibboleth that spending more money on technology is the best way to get better at doing business. The idea being that naturally spending more on technology enables higher automation and better maintenance and seamlessly transition to higher profitability. This also doesn't seem to hold true, funnily enough, because spending more money quite clearly doesn't translate into profitability!

But sine qua non making things work better is important. Maintenance increases a system's efficiency. You can do more with less, or at least make more with the same. There are legitimate arguments about where governments should play a role in this ecosystem, but there's definitely a role for a party that is a) interested in the long term growth, b) understands the "insurance-like" nature of the problem, and c) can bring enough resources to solve a market failure - whether it's brought about by tragedy of the commons of just lack of spare productive capacity.

It's fascinating as well, in this light, to view the fight between the politically polarised factions, which seems to often come down to the fact that conservatives feel the government doesn't function very well (which is true), whereas the liberals feel the government should do more for its citizens (which is also true). The funny thing is that these are orthogonal attitudes, rather than directly competitive. And that's also why a large number of the direct critiques on either side go unaddressed. Both sidestep the fact that things aren't maintained well, one by admonishing and trying to demolish it, and the other by trying to build our way out of the hole.

The flip side of the maintenance conundrum is the way budgets are allocated in companies today. With all the individual enthusiasm and incentives aligned towards building things, one would think that the part most in charge of innovation, the R&D team, would be laser focused on building new things. When they fall short, it would be because they've overextended themselves. I remember when I was in school, cold fusion was around the corner. When my dad was in school, this was flying cars powered by atomic energy. Whether we spent our budget wisely or not, it is safe to conclude that these dreams were tempered (rather fiercely) by reality. The reality of innovation was far more prosaic.

But if you look at the line items that most companies spend on, a majority of the spend, 60% to 80%, goes in what's called the "run" bucket, which is the part of the budget charged with keeping the lights on. In comparison the "grow and transform" areas, which are respectively meant to help with growth initiatives aimed at changing and growing the company, are usually the minority.

Is it just in business? Doesn't seem to be the case. It's also visible in Higher Ed.

This seems to be the case even in innovative companies. While it does bring up the fact that innovative endeavours have a high level of impact compared to the investment, its still incredible that most companies focus most of their IT investment in "maintenance" expenditures.

In fact, a study by McKinsey found that the average correlation of budgets from one year to the next was 0.92. A third of the companies had a correlation of 0.99, and while the highest degree of budget delta was 80%. This is remarkable, because it indicates that the capital allocation in one year is almost exactly the same as the year before, and even the companies that try to do a better job of re-evaluating their spending habits. This is also despite the fact that the same study showed that companies that reallocate their capital more year over year had an extra 2.4% annualised growth rate over a fifteen year period.

Without spending on maintenance, things fall apart. Keep spending more and more on it, things ossify. That's the paradox of innovation, that it only works at the margins once you have something to build off of.

Part of the answer is that when companies spend most of their dollars into making sure the lights stay on, the ability to change much of it on an ongoing basis will be limited. Let's look at an example, real estate, how I manage my home. Regardless of most circumstances, I wouldn't stop paying for the upkeep of my house, regardless of other priorities, and upgrading my kitchen and putting in a conservatory would very much be in the "nice to have" category. As a result, the correlation of my yearly spend on my house is likely to be extremely high. But that's also because my house has little chance of expanding from its current configuration into a mansion. That's just not how houses work. A business however, is not constrained by its physical size and scope. It can expand. It can grow. Even in the absence of clear investment into "building" activities, there is some inherent expansion built into its DNA.

But this is the status quo. Companies and governments and entities of all sizes invest more and more of their resources in order to preserve and protect what they have. What already exists. It is the most literal form of conservatism that exists. Not to say it's a waste, because it is essential. After all, ensuring the continuance of what's existing is the only way we get to build something new.

What used to take months of coding twenty years ago, to build a website or a simple application, now takes hours. That's what building and maintaining great tools gets you. You pay a bit more in maintenance costs, either insourced or outsourced, which is the complexity premium that comes with standing atop a prebuilt mountain, but without which nothing remotely meaningful can be built.

We need it to ensure what's been built stays built. Entropy is a powerful force, and counteracting it takes effort. But it's only if we counter it strongly enough and for long enough that we are able to build anything for the long term.

In fact, all progress depends on our ability to continually maintain and upkeep what we have, even as we add our own inventions and improve the way the world works. We have to learn to keep what we have already made to be able to make new things.

III

One could argue that my profession, in venture capital, exists to disrupt this status quo. To identify ways in which disruptive technology innovation can shake up entire industries and change the way we live and perceive the world. That's how we perceive ourselves, as guides ushering in a new era, and of confidantes to the brave entrepreneurs who do the work of ushering it in.

We want to be the ones to help build the anti gravity device. A large percentage of the people were directly influenced by the science fiction tropes by Asimov and Clarke, with flying cars and teleporters and faster than light travel and giant transporters wide in our imagination. We want to invest in and work with Elon Musks, and go to space and build cars.

But those are hard goals to achieve. Turns out sending hunks of metal whizzing around, or into orbit, isn't as straightforward as of looks. And as much as we like to think of ourselves this way, funnily enough, a large part of the theses that most VCs (like me) make also revolve around efficiency. It gets muddied because this isn't explicitly in either camp of building or maintenance. Oftentimes it's a hybrid of both. Where would Uber lie? It didn't particularly build something new, but helped dramatically improve on the utilisation and distribution of the assets.

In many ways, we fail the 'atom test', the identification of ways in which our endeavours change how you react and deal with the physical world surrounding you. Whether there is a shift in how the molecules touch your body.

There are occasional forays into areas full of untapped possibility of course, of building a new industry perhaps, as in the case of biotech, clean tech, climate tech, to name a few. Current favourites include quantum computing too, in that series. There are specialised investors who look at these spaces for sure, and some who have made great trades, but it isn't an area that was known for extraordinary returns.

Instead all the focus is on efficiency-hunters. Those companies who manage to identify a niche that was not particularly well served, and create products that would solve that niche. Oh, developers are spending time searching for pieces of code on Google? Boom, Github. Oh, office workers spend hours copy pasting pieces of data into excel? Boom, UIPath and RPA. The promise of all of these pieces of individual 'efficiency innovation' is that they remove drudgery from our daily jobs. And once you remove drudgery, you free up people to be their best selves, innovate more, ideate more, create more, solve all of the world's problems.

That's the dream anyway.

What about coming up with better ways for us to communicate, and free us from email? Slack! Did that solve a problem or just carve out a niche and apportion part of the gains from one system to another? What about better productivity tools to help organise the clutter of our minds? Asana! Complete with Buddhist practices and an engaged CEO. Did that solve our critical problem of coordination? Not yet, I would say.

The rhetoric around innovation has mostly been driven by companies which are either reallocating resources through technology and reaping rewards of being that hub of reallocation (hello again, Uber), or those who create new business model combining shiny new technology (GPT-3, Tensorflow) and a clear way to attract users. It's a miracle of modern accounting. If you can save 2% of effort for 1000s of people, they will switch their providers to use you instead. The most drastic example might be Superhuman, reinventing the 2 decades old phenomenon of email, by promising to save its users fractions of an hour each day. The key claim was that everything inside the app happened within 100ms or less, and provides keyboard shortcuts to make things even faster. That's the hyper-sliced sliver of time that we are attempting to improve upon now.

It's not to say it doesn't improve our lives. Or even that it's necessarily net neutral in the effort to build something and sell it vs not do that. Sometimes doing ten neutral things is needed to have a platform for the eleventh to break through. Like Tetris, where you need to pile up a bunch of misshapen objects just so, not because it's a better configuration, but because it prepares the field for that eleventh piece that makes three entire rows disappear.

And what happens to the 'masons' of our age, as this progresses? The slivers of efficiency gain get funneled to the monument making Kings (the companies) and the efficiency creating inventors. The formerly skilled mason is lucky if she's not fed to the crocodile.

The world is very different for sure in this iteration. Things that took months take hours. The man-hours required to build our monuments, physical and mental, are dramatically lower. A postcard has the same processing power that once took humanity to the moon. And yet, and yet, the masons still remain under threat. They become the disposable parts of the system, as our visibility of what actually drives our monument creation is obscured.

But perhaps the worst part of that those masons who are skilled enough to think of something anew, braving the wrath of the Inventors, they're dwindling too. Unless they King takes pity on her, or brings her to the Inventors fold, crocodiles or ignominy it is.

And there we are, back at the original conundrum. How much should we spend on maintenance, and how much on building. Sometimes the right answer is that optimisation is essential, hunt for efficiency is crucial, the ROI hunters like us VCs can be gotten excited, and we can free up ourselves from the drudgery of trying to save each other man-hours. But it requires us to be aware that within a bounded system, increasing that efficiency only increases the displacement the people in the system feel. Until the anti-gravity device is in place, we'll all be stuck in an inadequate equilibrium. And the only way out is to employ our best and brightest minds to find the next breakthrough. And that's not inevitable, neither is it unachievable, but it's just something we need to aspire to.

Progress and innovation depends on standing on a firm footing. It requires maintenance, not just an obsession with building, and without it we are just mechanising parts of the labour net, without necessarily adding much empowerment along the way. Just as we can acknowledge that the invisible hand has guided us towards focusing on optimisation puzzles and not a paradigm shift, we have to admit that this isn't a process where the outcome is always visible. In fact it rarely is. While we shuffle the deck chairs and eke out the gains, it is important that we put resources and significant effort into utilising that spare capacity that's being freed up. As we drive the maintenance cost to be lower, we have to work hard to ensure we are spending that freedom well. That's the secret of all progress. That's how we see the change happening in the world gradually, and then suddenly.