AWS Cloud Enterprise Strategy Blog

Buy vs. Build Revisited, Part 2: Drawing the Line

My previous post highlighted that even seemingly straightforward decisions can become challenging in the context of the vast scope and complexity of enterprise IT. For example, when considering whether to build a bespoke software in house over buying a commercial one, many companies overlook several critical nuances:

  • You pay in opportunity cost, not just direct cost
  • Value delivery hinges on an end-to-end value chain, not just building applications
  • You’re locked into a built solution just as you would with a purchased one

Things are apparently never simple in enterprise IT, so let’s see how we can do better.

Heuristics for Better Decisions

Direction sign concept.Enterprises can’t afford to fall into analysis paralysis every time they need to choose left or right. At the same time, with large sums of money involved, they also don’t want to just follow their gut instinct. Experienced executives and architects therefore resort to heuristics, which are “simple strategies or mental processes that humans, animals, organizations, and machines use to quickly form judgments, make decisions, and find solutions to complex problems” (Wikipedia). The heuristics we employ are typically based on a mix of past experience and good industry practices—i.e., experiences shared by others.

While inherently fuzzy, heuristics help IT executives and enterprise architects keep their sanity by making meaningful decisions under complexity, uncertainty, and time pressure. Heuristics often employ mental models, which help us navigate uncertainty and rapid change. For example, we frequently use mental models here at Amazon to help us make better decisions.

Naturally, even the best heuristics need to be periodically updated as technical capabilities and constraints shift. “Just put the giant monolith on the mainframe” might be a quick decision, but likely not one that matches this decade’s IT environment.

What Differentiates? You Choose!

Buy vs. build decisions for systems are frequent enough that many enterprises rely on a common heuristic:

If the functionality differentiates you, build the solution in house; otherwise,
purchase commercial off-the-shelf software.

A chef puts the finishing touches on a plateSimple enough! Now all we need to do is decide whether the needed system would differentiate our business and choose accordingly. The heuristic is easily applied: employees likely just want to get paid the correct amount, making payroll systems, like many other HR systems, easy candidates for buying a solution. The same goes for most financial or resource planning applications. Looking outward from enterprise software into the physical world, most facilities such as restrooms and cafeterias are going to be easy decisions.

Hold on a second, though…don’t the Silicon Valley companies, the ones seen as the epitome of modern, agile, and digital organizations, routinely advertise their wide range of healthy and organic food prepared by full-time celebrity chefs in their fancy cafeterias rebranded as cafes? Perhaps they decided to “build” their cafeteria menu just because they had money to give away? Not at all!

These companies recognized two important factors that influence their buy-vs-build decision:

  1. Differentiation isn’t just directed toward your customers, but also to employees, partners, suppliers, investors, and other constituents
  2. Any commoditized offering is also an opportunity for differentiation

So, when the ability to attract top talent became a limiting factor for these organizations, they decided to turn something that wasn’t a differentiator into one: better (and free) food. Flipping from buying to building had a significant upside for the growth of their business, at least for a while—more on that later.

Since then, services ranging from onsite haircuts and car washes to free snacks, gyms, yoga courses, massages, and a whole litany of other ancillary services have become differentiators for high-tech companies to the extent that a free barista stand has become the de facto indicator for being a “digital” organization, or at least management’s desire to appear like one.

Enterprise Differentiators

Serving organic food in the cafeteria is nice and all, but what about major critical IT infrastructure investments that can reach into the tens or sometimes even hundreds of millions of dollars? Should a bank or insurance company build its own core system? Should a retail business build its own cash register? What about manufacturing and supply chain management software? The answer invariably leads to the dreaded “it depends,” so let’s have a think about what it depends on. We’ll stay with the theme of food, but we trade the employee cafeterias, sorry, cafes, for a venerable food & beverage business.

Registering Cash

TV remote control in the foreground, tv in the blurry backgroundMy Enterprise Strategy colleague Phil Le-Brun led several critical buy-vs-build decisions as Corporate Vice President and International CIO at McDonald’s Corporation. As a quick service food and beverage business, which offers speedy service and consistent quality, McDonald’s naturally requires a system for customers to order and pay: a cash register or POS (point of sale) in modern parlance. Given that customers come for the food and the service, McDonald’s surely wouldn’t want to build their own ordering terminal. Or would they?

After attempts at procuring off-the-shelf POS software, McDonald’s came to realize that their POS is much more than a simple order entry and payment device. It’s really the operating system for the restaurant and a differentiator for both operational efficiency and customer experience. They therefore decided to build their own integrated system.

The benefits became even more pronounced as consumer expectations started to shift toward loyalty programs, new marketing campaigns, preordering, and omnichannel engagement. Using a commercial platform would have meant waiting for a vendor to see sufficient demand pent-up across the industry, eradicating any possible differentiation.

Selling Content

TV remote control in the foreground, tv in the blurry backgroundSimilarly, my colleague Ishit Vachhrajani went through several major buy-vs-build decisions as CTO of A+E Networks. Content sales systems are an integral part of the media and entertainment industry with several commercial solutions available in the market, which cover the typical sales pipeline, including deal and order management plus fulfillment.

However, A+E Networks realized that in addition to selling content in response to demand, they also wanted to proactively drive such demand. For example, they could use past consumer behavior to drive future consumer behavior. Packaged solutions could have led to a potentially messy integration, so Ishit and his team decided to build an end-to-end solution in house. The real payoff came when the industry shifted from syndicated content sales to video on demand. Having control over their system allowed them to support a new line of business in a matter weeks (please see Ishit’s related blog post for more insights).

Where to Draw the Line

A finger draws a line in the sandThese examples highlight that drawing the line between when it’s good to build and when it’s good to buy isn’t easy after all. They also remind us that good IT decisions depend on good business decisions.

Here are a few simple rules (dare we say “heuristics”?) that can help you draw the line between when it’s good to buy and when it’s better to build.


1. Know Thy Business

McDonald’s decided to build its own POS because it saw its POS as more than an order-taking device. Its restaurants are effectively retail sites coupled with a manufacturing back end, multiple delivery channels (drive-thru, counter, delivery, table service), and multiple order channels (mobile, counter, kiosk). Having these functionalities integrated into a system formed the basis for competitive activities such as successful pricing and marketing strategies.

Once you thoroughly understand the business problem, the IT decision becomes apparent.

There’s a piece of old wisdom commonly attributed to Einstein: “If I had one hour to solve a problem, I would spend 55 minutes thinking about the problem and 5 minutes thinking about solutions.” This applies to IT more than anywhere else. Once you thoroughly understand the business problem, the IT decision quickly becomes apparent.

2. Don’t Rebuild What Is Already Available

The corollary to the prior point is that building a near copy of an existing system is unlikely to be a useful endeavor. Companies that rebuilt a POS or a content sales system had a very different idea of what such a system should comprise. Therefore, buy vs. build is much more than taking one parcel from your IT landscape and filling it with either a commercial system or a bespoke one. Rather, it’s a matter of rearranging your IT landscape to match the business strategy and then building something that matches this new landscape.

Buy vs. build isn’t about taking one parcel from your IT landscape and filling it with either a commercial system or a bespoke one. It’s about rearranging your landscape.

3. Velocity Favors Building

In both examples cited above, the custom solution really paid off when the industry shifted to a different business or engagement model. Using a commercial solution, you’d have to wait for the vendor to deliver the desired feature. In a slow-moving industry, waiting six to twelve months for a solution likely isn’t so bad. In a fast-moving world governed by economies of speed, however, a six-month lead over the competition can significantly shift the playing field in your favor. That’s why you see so many digital companies build solutions despite their high opportunity cost.

Don’t consider just the features needed today. Also consider the degree of uncertainty: in other words, the likelihood that new requirements will arise due to competitive pressures or technology evolution.

4. It Won’t Last

The sad news is that even if you were able to draw the correct dividing line between “buy” situations and “build” situations, or differentiating and non-differentiating capabilities, the boundary isn’t static. Today’s differentiator is tomorrow’s commodity, as evidenced by the abundance of gourmet cafes in Silicon Valley. My colleague Mark Schwartz captured this effect in his aptly title blog post Unsustainable Competitive Advantage. And vice versa, today’s commodities can become the base for tomorrow’s differentiators. You’ll want to reserve some of your delivery capacity for new ideas or sudden market shifts. 2020 is going to be a stark reminder for how quickly such shifts can take place.

Next Time: Transitioning From a Bought System to a Built System

Assume you’ve correctly concluded that despite the high opportunity cost, you’re best advised to build a custom solution. Likely, you won’t just replace the existing system 1:1 with yours; that would defeat much of the purpose of building in the first place. As you might have guessed, those decisions also involve important nuances. We’ll dive into those in the next post.

Twitter | LinkedIn | Blogs | Email


More on this topic

Buy vs. Build Revisited: 3 Traps to Avoid, Gregor Hohpe

Don’t Outsource Thinking, Gregor Hohpe

Time to Rethink Build vs Buy, Ishit Vachhrajani

Gregor Hohpe

Gregor Hohpe

In his role as Enterprise Strategist at Amazon Web Services, Gregor advises technology leaders in the transformation of both their technology platform and their organization. Drawing on his experience as Smart Nation Fellow to the Singapore government and as Chief Architect at Allianz SE, he connects the corporate strategy with technical decision making and vice versa. He enjoys sharing his thoughts on architecture and architects in his books, including The Software Architect Elevator and Cloud Strategy.