Sunday Reads #144: "Sell the mailroom!"
Peter Drucker's 1989 prediction and the strange way it's coming true.
Hope you’re having a good weekend. And happy Chinese New Year, if you celebrate.
Earlier this week, I re-read one of Peter Drucker's most famous essays - Sell the Mailroom. It appeared in the Wall Street Journal in July 1989, and caused quite a stir.
It resonated so much that WSJ reprinted it 16 years later to commemorate Drucker, when he died in 2005.
The basic thesis of the article is this: corporations will be unbundled.
More and more people working in and for organizations will actually be on the payroll of an independent outside contractor.
Businesses, hospitals, schools, governments, labor unions -- all kinds of organizations, large and small -- are increasingly "unbundling" clerical, maintenance and support work
It’s interesting to re-read this article in 2022, and see how it's aged. What it got wrong, what it got right, and where reality is even more radical than the prediction.
Drucker cites three key reasons why an unbundling is inevitable.
#1. Support work is costly to do
Support work is rapidly becoming capital-intensive…
In many manufacturing companies, the investment in information technology for each office employee now equals the investment in machinery for each production worker. Yet the productivity of clerical, maintenance and support work is dismally low, and is improving only at snail's pace, if at all…
Forty years ago, service and support costs accounted for no more than 10% or 15% of total costs. So long as they were so marginal, their low productivity did not matter. Now that they are more likely to take 40 cents out of every dollar they can no longer be brushed aside.
This was undoubtedly true in the 70s and 80s.
I've just finished reading Becoming Trader Joe, a founding story of the value retail chain.
Joe Coulombe, the founder of Trader Joe's, did exactly what Drucker talks about.
He focused the entire organization on its core value proposition as a retailer: buying and selling. He outsourced everything else.
IT and computing → outsourced
Maintenance → outsourced
Real Estate → outsourced
Result: Revenue quadrupled between 1977-1987, but profits grew 6x-9x!
Now here's the thing though - Drucker's observation is not necessarily true today. One thing he did not predict, was the steep fall in the cost of computing.
As PCs became more powerful, the capex around many support functions almost vanished. And many activities got automated. This reduced costs and increased productivity even more.
Of course, options to outsource non-core functions are more plentiful than ever. Some of you will read this essay at a WeWork. Most large offices use facility managers.
But it's not because of cost. It's because of the hassle of managing non-core functions.
Prediction Score: 6/10
#2. In-house support functions have no incentive to be productive.
In-house service and support activities are de facto monopolies. They have little incentive to improve their productivity. There is, after all, no competition…
When in-house support staff are criticized for doing a poor job, their managers are likely to respond by hiring more people.
An outside contractor knows that he will be tossed out and replaced by a better-performing competitor unless he improves quality and cuts costs.
This is a great point. Incentives matter. Or as Charlie Munger would say:
But the Internet has reduced the monopoly power of support functions quite a bit.
It's now easy for any CEO to estimate the typical support costs for an organization their size. Just need to type the question into Google. Or download the Annual Report of a public company in the same space.
What can be measured tends to get managed, after all.
Having said that, Drucker's point is still true at the margin. You can still make a support function more productive by outsourcing it to an expert.
Aside: While we’re on the subject, there’s one more way to increase the productivity of a support function. The Amazon way.
Amazon took each cost center (server space, warehousing, logistics, etc.) and made it a revenue center. They offered each of these as a service to external customers (AWS, Fulfilled by Amazon, Amazon Logistics, etc.).
Now, the different functions actually needed to sell their services. They all had P&L targets to hit. They suddenly had to compete with other players, and win.
Competition is a great way to get the best out of a player. (I'm writing this as Nadal attempts to go past Djokovic and Federer at the Australian Open final).
Prediction Score: 7/10
#3. There are no career growth incentives in support functions.
The productivity of support work is not likely to go up until it is possible to be promoted into senior management for doing a good job at it. And that will happen in support work only when such work is done by separate, free-standing enterprises…
As employees of a college, managers of student dining will never be anything but subordinates. In an independent catering company they can rise to be vice president in charge of feeding the students in a dozen schools; they might even become CEOs of their firms.
Drucker is bang-on here. Career growth is a powerful motivator to work hard and improve productivity. Without that, organizations would see dwindling productivity and / or high attrition.
The latter is what’s happening today.
Organizations do manage some support functions in-house. But these departments tend to have high attrition. The only way to get career growth is to jump to a new company.
Prediction Score: 7/10.
Overall, I'd rate the article a solid 7/10 on how well it has aged.
But there's at least one way in which this unbundling is playing out, which Drucker could not have predicted…
The Era of Radical Unbundling.
When we hear "support function", we think of legal, accounting, facility management, etc.
But, if you step back, the entire management cadre of any organization is a support function!
It is a necessary support function, of course. There's no other way to align the entire organization in a common direction.
Or is there?
Coz unless you're new here, that's exactly what blockchains are supposed to do!
I wrote about this in The future of the workplace (it might not exist in 2030):
Why do firms exist at all?
Organization structures are one way to reduce transaction costs.
Another is technology.
That's where blockchains come in. Blockchains and smart contracts…
[Today], if you accept payments using Stripe, you pay fees to Stripe. Not to the protocol. You pay Google for storage. You pay Facebook (with your attention).
The blockchain changes all this. You don't pay any company for using their service.
You pay the blockchain.
And the blockchain automatically pays agents that provide it a service (e.g., token miners, validators, etc.). And yes, all the rules are right there - encoded in the chain for all to see.
No adult supervision needed.
I’ve given an example of this before. Etherisc: A decentralized insurance protocol that replaces an entire insurance organization.
Won’t go into the details (check out the thread 👆), but here’s a visualization of how the Etherisc protocol works with different parts of the insurance ecosystem.
We may be in the Roaring ‘20s, but the ‘30s will be a strange time.
Golden Nugget of the Week.
My new favorite term is Pigeon Chess. I've gotten quite good at playing pigeon in my Twitter conversations, if I say so myself 😉.
This made me laugh out loud 🤣
That’s it for this week. As always, stay safe, healthy, and sane.
PS. I have a question for you. There’s a lot of incredible business content on Twitter - informative, inspiring, and often hilarious too! Would you like a “Best of Twitter” edition of Sunday Reads every once in a while? Hit reply and let me know!