The Best System Design Doesn't Always Win
A while back, another engineer and I spent an afternoon talking through two plans for the same piece of database architecture.
We were deciding which object should own the relationship with a connected account. Our designs were structurally different, and either one would've solved the problem. Every point one of us raised had a reasonable answer from the other.
Mine got the nod. Looking back, the decision was fine, but it wasn't any better either way. It moved forward because I talked about it in terms of what it meant for our customers.
What "owning the relationship" means: in a data model, the owning object is the one that holds the reference to the other. That choice shapes what has to exist before two records can be connected, and what happens to one when the other is removed. |
Focus on the Business Outcome
I talked through my approach from the customer's perspective. It required fewer steps to connect an account, and it simplified the flow a customer had to go through to get there.
That's what got traction. The people making the call cared about how quickly customers could sign up and connect, and they could see how this design helped.
The discussion about the other design stayed mostly on the technical details. It may have had benefits for customers too, but they never made it into the conversation, so there was nothing on that side for the business to weigh.
The technical details still mattered. Both designs had to hold up before either one could be considered. Once they did, the customer experience was what separated them.
Your Coworkers Are Listening for Impact
When you bring a technical decision to people outside your immediate team, they usually aren't evaluating the design itself. They're asking what it means for the things they're responsible for.
- A product manager wants to know how the change affects the customer experience and the roadmap.
- An engineering leader wants to know what it takes to build and maintain, and what it puts at risk.
- Sales and customer success want to know whether it makes the product easier to buy, adopt, or support.
If your explanation stops at the implementation, they have to translate it themselves. Some will, but others will lean toward whichever option someone already explained in their terms.
This applies to far more than architecture debates. A request to refactor the billing module lands differently when it's framed as "every pricing change takes us two sprints right now." It's part of understanding all the systems you work with, including the business ones.
|
|
Where can AI save you time?
My friends at Big Creek Growth put together a quick survey to spot the repetitive work you can hand off to automation.
|
|
Find Out What the Business Is Already Measuring
You can't frame a design around business outcomes if you don't know which metrics are important to the business. That context rarely shows up in a ticket, so you have to find it.
A few places to start:
- Ask your product manager which metric the project is supposed to move. If there isn't one, that's useful to know before you start designing.
- Read your team's or company's goals for the quarter. They tell you what leadership will be listening for in any decision.
- Spend time in support tickets or sit in on a customer call. You'll hear where customers get stuck, in their own words.
- Ask the person making the call what a good outcome looks like to them before you present.
In my case, the business already cared about how quickly customers got signed up and connected. Knowing that made it clear which part of my design was worth talking about.
Frame Each Option Around Its Business Outcome
Before you bring a design decision to the wider team, answer these questions for each option:
- Who is impacted by this change? Customers, a support team, another engineering team, or whoever owns the cloud bill.
- How does their experience change? Fewer steps, fewer support tickets, faster onboarding, or a lower cost.
- How would you measure it? Name the number that should move, like time to connect an account, and know where it sits today.
- What does it cost, and how hard is it to undo? Some designs are cheap now and expensive later, and some are painful to reverse.
Take the design from that afternoon. Described technically, my approach sounded something like this:
Because the owner object owns the relationship, we don't need to create and configure anything else before we link the connected account.
Described by its impact on customers:
A customer can connect their account without setting anything else up first. It's fewer steps and a simpler flow during signup.
Both describe the same design. The second one is what gave the people making the decision something to weigh, and it did more for the outcome than the afternoon of schema discussion before it.
If you can't answer these questions for your own design, work them out before the meeting. You may find the options are closer than the debate made them feel, which is worth knowing too.
When Both Designs Work, the Business Outcome Sets Yours Apart
Two capable engineers will often land on two sound designs. When neither has a clear technical edge, the business outcome is what differentiates them.
That puts the engineer who can explain who benefits, and how, in a strong position. Connecting a design to customers and the business gives the team something concrete to decide on when the technical back-and-forth has stalled.
The habit also builds over time. Engineers who consistently tie their work to business outcomes tend to get pulled into decisions earlier, while there's still room to shape them.
If you lead a team, you can build this habit directly. When someone brings you two viable options, ask who the change impacts and how. Over time, your engineers will start showing up with those answers ready.
The next time a design discussion runs long, check whether the options really differ on technical merit. If they don't, "who is impacted by this change?" is a good question to move it forward.