The key rotates slower than usual
You hear a big thud behind you. The big entrance door just closed, and it bumped you to make another step. Based on the lights and sounds, your roommates are all at home. So, let’s see what they are doing.
The kitchen, located first on the right, exudes a lively energy. Upon entering, you are greeted by a sweet, summery scent—fresh-cut peaches emanating their subtle aroma, creating a welcoming ambiance reminiscent of sunlight trapped in the air. With a few more steps, the delightful peach fragrance is soon overshadowed by the robust smells of fresh tomatoes and warm dough.
Angelus slices and skins peaches with fury, trained hands moving like lightning, filling his tray with slice after slice. The freezer is open near him, maybe to combat the heat from the cooker, which was all emptied this morning. He already imagines tomorrow, opening a small bag of frozen fruit to translate into a smoothie. The freezer will be thoroughly cleaned in one month, and the cycle will continue.
You can taste the difference!
Sentia is sitting on fresh pizza dough. She confidently and meticulously ensures the correct amount of homemade tomato soup, buffalo mozzarella from a local farmer’s market, and basil leaves from the back garden. The extra virgin olive oil adds flavor and humidity later in the cooking process.
I don’t understand why people pay for this at a restaurant. It’s far more affordable to make it yourself at home.
Exiting the kitchen, you eye a big stack of nicely washed and folded clothes. They are stacked on top of one another with a string outlining them, making them easy to pick and open. Two of these stacks are linked with a ribbon spelling “Nepia,” the roommate nicknamed the lazy one. Less time doing laundry means more time writing code, something a software engineer should focus on.
But it’s only 30 bucks. Give me a break!
Getting closer to the living room, you already know who is connected to the soundbar: Palinurus. Suddenly, “Watson not responding to ping” flashed on the TV, and Beethoven was no longer playing in our ears. The terminal was opened, commands were given, dashboards were checked, and everything returned to normal after some minutes. Hopefully, only a few customers noticed the downtime. He fixed it fast, having experienced this already too many times to require writing the fix. Running on his servers is worth it, and no amount of mentors suggesting AWS would change his mind. Already in the fourth year of profitability, his five-person company has only enjoyed the benefits of keeping server costs low—no need for investors to take his flexibility and independence.
Why to do that? Why limit on AWS when I can run it everywhere?
And you arrive at your desk, trying to reflect on this small encounter. Pen and paper are your allies in this adventure.
Is it that simple?
We are not adding value to this, so we should outsource it.
We are not in the business of doing that. We should not reinvent the wheel and let the companies in this domain run it for us.
The four characters above represent differing perspectives on outsourcing tasks. Previously, I viewed the choice as straightforward: if your time is precious, follow Nepia’s example and delegate or outsource as much as possible. Failing to do so seemed like an unreasonable loss. I quietly critiqued those who influenced Sentia and Angelus.
So, let me share some of my thoughts regarding outsourcing (or delegating) to a person or company. Currently, I pay for many things most people do themselves, and after making this decision several times, I can share some less-known factors and observations about the process. Sometimes, the quality is way better if an expert company with specialized tools handles the problem, thus making it cheaper for you in the long term. Sometimes, it can be more affordable to do it yourself, even if you value your time highly and the other person is doing it much faster.
An old problem, highly discussed
Many people have discussed this on the Internet. A quick search of “buy versus build” will lead you down a rabbit hole. So, let’s see what I could see through the cracks.
Some articles are written from the corporate boardroom, others from the home kitchen. Almost all conclude that it’s a choice between spending time (or “in-house resources”) or money for a similar value. Immersing ourselves further in the rabbit hole, we will find sophisticated articles outlining transaction costs (if you don’t want to teach your kid something, you still need to find a tutor) or principal-agent problems (needing to vet that tutor).
Observation one: Build-or-buy decisions are not about two alternative ways of getting the same thing but about two options sufficiently different acting as separate objects with overlapping benefits.
Looking at one example, it’s straightforward for a home baker to explain why a supermarket can’t compete with homemade bread. Thus, the decision is easy. But we tend to prefer models of the world, so let me try to suggest one that can help make any decision easier.
So what do we want?
A quick search on Etsy or Instructables reveals that the supply of homemade items is vastly more significant than that of store-bought items. But do these satisfy the same demand?
It would be unwise to think that providers don’t try to fulfill your needs (since this is their source of income). Yet outsourced providers balance the needs of many customers and face problems you don’t. Their supply may be meaningfully equivalent, sometimes for the worst, even in the face of competition. These differences encompass the product or service and the accompanying intangibles.
You are the only person who wants the exact things you want. So, observation two: when you buy something, some of its cost goes into things you do not want.
Providers need to tailor their offerings to attract a wide range of consumers. If your preferences vary from the average, you end up paying for products that aren’t well-suited to you. No one is close to the median in everything, and plenty of people go through phases of unusual preferences.
Outside of arbitrary tastes, your distinct preferences can also come from your position in life. For instance, common market wisdom is that it’s almost impossible for an individual to beat the market index funds. But an index fund is an off-the-shelf product, selling a standard risk package. You want a package of risk that anti-correlates to your career. So, an American software engineer can “beat the market” to maximize their future wealth simply by not buying US tech stocks.
Off-the-shelf offerings are designed by distilling the preferences of a large group of people into a few simple components. In contrast, you can craft a highly complex deal with yourself, containing exactly what you want and no more.
Accommodating many preferences often comes in the form of bundles. So, those who use less pay for those who use more. If your plan includes unlimited customer support, you subsidize the most demanding customers. Similarly, if you fly an airline and do not use your free checked bag, you subsidize everyone who does.
So what is my guarantee?
Observation three: This is also present in physical products. When you buy a product, you are also paying for the risk of the average user.
Do you know the coffee cups’ label “Caution: hot product”? This comes from the modern doctrine of strict liability: companies are, by default, liable for damages caused by their products, even without identified fault.
And the easiest way to provide “insurance” to their customers is by raising the price tag. Some portion of every mass-market product you buy is paying for potential lawsuits that result from that product. This is usually a good thing and a reason to outsource, as you’ll be in much better shape financially if your house gets burnt down by a malfunctioning toaster than your hand-sculpted wood-burning oven. But this also means that even if your home is fireproof, you still pay for the risk because someone else isn’t. If you buy a chair, you’re paying for the potential injuries because somebody else assembled it incorrectly. In general, if you think you’ll be better informed or more careful than the average consumer, then you are losing out.
Toasters don’t often burn homes, and so this cost is low for most products. But it’s still present and more substantial in, of course, actual insurance. When someone with good oral hygiene chooses to forgo dental insurance, they are basically in-sourcing.
Several other factors here have a common theme: you pay more because of information you don’t have about the product or service. This one is the reverse…
Observation four: you pay more because of what the seller doesn’t know about you.
Technique
In physical products, centralized production, distribution, and storage constraints may cause specialized producers to use techniques different from those of DIYers. This can sometimes result in an inferior or superior product.
For example, Angelus in the kitchen has only disadvantages compared to commercial producers of frozen fruit. He doesn’t have a blast freezer (which results in better taste and higher speed), and the producer's central location gives them access to fresher ingredients than him.
In contrast, store-bought tomato sauce is also different from the homemade version, but in a negative way. The tomatoes are peeled either by immersion in an acidic solution (requiring dye to compensate for the loss of color) or by scalding and physical trauma until the skin falls off. The long storage times require either preservatives or pasteurization. Centralization means scale, and scale creates its problems. Compared to Sentia, the tomato sauce manufacturer pays for scale by degrading health and flavor.
Even for virtual products, centralization makes a big difference when shared physical resources are used to satisfy many customers. For example, AWS faces the risk that customers with VMs on the same machine may attempt to snoop on each other. They pay for this by degrading performance with expensive mitigations against information-flow attacks and wiping disks each time they’re connected to a newly booted VM. If you own those physical services yourself, you might consider these to be unnecessary.
And yes, this cost is usually attributed to something we call overhead. When you buy frozen fruit, you also pay for the transport chain, shelf space, and the cubicles of designers who create packaging and advertising—and, of course, subsidize the unsold product. When you buy enterprise software, you also help pay for the vendor’s sales team, HR, and finance departments, ensuring the presence of the pricing dead zone between four- and five-figure contracts.
Principal/Agent Problems
Delegation is a coordination problem, and solving it means solving principal/agent problems: communication and risk. I am sure you’ve heard stories of hiring an employee who turned out to suck up more manager time versus if the manager had done it themselves. It can be whether the hiree couldn’t deliver (risk) or required too much help to understand the task (communication). This is not a new topic; it is widely understood, although it is often underestimated by the less experienced.
The need for signaling costs
When I cook a roast for myself, I might toss it in the oven and return a few hours later without setting a timer or checking the thermometer. It might taste worse, but this doesn’t tell me whether I can make a nice steak when I want to.
Now imagine a restaurant that sold a six euro roast, prepared as cheaply as possible and a fifty euro steak. A consumer of the first is less likely to conclude, “The roast is worth 5 euros,” and more likely to say, “The food here sucks.” This makes it so every provider must invest in signaling mechanisms that showcase quality without direct contribution. Cognitive fallacies, which companies exploit in areas such as “dark patterns,” can provide examples of this.
The reverse is more subtle.
Observation five: every provider must avoid visibly cutting corners, even when doing so would benefit the customer.
When organizing events some time back, I explicitly told a cleaning company that I didn’t care about a specific area. When I was notified that the cleaning was done, I checked that unimportant area, which was spotless. This was not a planned “test,” but I kept the same company for multiple events. This happens quite often in academic work: a peer reviewer with illegitimate criticism requests that the paper be published by performing work of questionable utility.
This fun is unnecessary if the consumer and producer have enough shared knowledge. This also means that when buying on the mass market, you are paying for a presentation, and the cheapest option is usually hidden and accessible only with the correct knowledge.
With shared knowledge comes great opportunities
An extra benefit to outsourcing is that the information can be reused in future transactions. After a professional has proven their quality, you also buy a relationship.
When looking for a cleaning company for big events, I asked friends, posted an ad online, and tested a few competing companies. Now, all I do is send an email with the location, date, and surface area and receive an invoice the following day.
Expertise and its costs
One way to determine what to delegate is to consider how much training it takes for someone else to do it. Based on this, you can decide whether to outsource or eliminate it if it’s too low.
Observation six: the higher your expertise, the more people you should collaborate with.
Well, skilled professionals routinely perform tasks beyond their training and charge full price. For example, an auto shop will bill you for a minimum of one hour of mechanic time for a tire change. On the other side, a high-end lawyer needs to do some low-level contract review in the middle of a more extensive engagement.
This happens because of bundling and coordination costs, but it also occurs because high-end sellers are incentivized to convince you that everything they do is worth the markup. These incentives create attributes like “natural,” “organic,” “premium,” and “value-added” that mostly mean nothing while increasing the margins.
These are also woven into our culture. Consider the tale of an engineering consultant who charges $10,000 for a chalk mark on a faulty coil: $1 for the mark itself and the remainder for knowing where to place it. Or think about Picasso charging 5,000 francs for a two-minute sketch, proclaiming, “It took me my whole life.” Would you notice the distinction if you paid for an answer derived from years of training versus one from just an hour of online searching?
Providers also distinguish their offerings from DIY versions by bundling added value. A cafe’s ambiance, an expert’s eye, or a thousand customers before you have had a great outcome. The sophisticated buyer can see when a bundle component isn’t pulling its weight and is prepared to craft its own to compete. A fool and his money are soon parted. The reverse is also true:
Observation seven: You will always find sellers trying to convince you that a $15 avocado toast is worth it. Wealth comes from becoming the sophisticated customer who says no.
The value of learning is on a spectrum
A paradox of outsourcing: the better you can do something yourself, the less valuable it is to do it yourself.
When you do something yourself, you always benefit from having it done and then learning how to do it. Doing laundry may be busywork to the habituated, but doing it for the first time also offers the security of not being dependent on others and being prepared to handle the situation where the laundromat returns wrinkled clothes, claiming, “That just happens.”
The value of learning differs significantly between individuals. My ambition in cooking is to impress guests a few times a year, so I’m not likely to benefit from handcrafting pizzas. Conversely, a French chef may learn valuable things about his suppliers from growing his vegetables or gain insights from branching out into other cuisines and making his own tortillas.
Should I buy or build?
Yeah, not so easy, right? I do have some guiding questions from all of this reading. These should help a bit make your decision more straightforward:
- Are there aspects I don’t care about and providers take great pains to provide?
- Does my preferred set of tradeoffs differ substantially from the average buyer?
- Do I believe I am savvier than the average consumer of this product or service?
– In particular, is there a considerable risk associated with this offering, and do I have reason to believe I am much less likely to experience it than average? - Do providers have access to specialized capital or technology for which I lack suitable substitutes?
- Do providers have substantial overhead that I would not?
- How much of the cost is spent on signaling quality?
– Do I expect to do business with the same provider again? - How much of the price are the transaction costs?
- How do I know whether and when the unique expertise claimed by the provider is beneficial?
- How much do I value the learning from building it myself? Do I expect to use the knowledge gained? Is the learning involved overly specialized?
- Are there other ways in which the thing to be bought substantially differs from what I’d do for myself?
