My wife is running the Chicago Marathon this weekend, so I'm wrapping up this article tens of thousands of feet in the air. How fun is that? It's a fantastic privilege to be able to spend our time on the things we value.
On the other hand, my chair at home is more comfortable than economy seats on an airplane. Actually, sitting on a wooden box in the rain would be more comfortable than sitting in this seat. At least I'd have room to stretch my legs.
But I digress. You can only enjoy luxury if you experience non-luxury. I will appreciate the rest of life that much more after being stuck in a small metal box for a few hours.
As happens occasionally, I ended up with a long list of topics I wanted to write about, but none of them felt like a whole article. Rather than write a short article, or pick a new topic, I thought I'd make that my theme. My theme is, "A laundry list of topics Dave wanted to write about!" It also became a long article. Whoops!
Sign up for Scarlet Ink by Dave Anderson
Weekly leadership advice from an ex-Amazon GM and Tech Director. Most articles have substantial previews available for free members. Paid members can read all current and future articles, and are statistically more awesome than the average person.
No spam. Unsubscribe is self-service at anytime.
These are topics from mentoring sessions, business collaborations, school meetings, discussions with my kids, and much more! How much fun. For me at least. Hopefully, for you as well.
Sit back in your chair, which is surely more comfortable than mine, and enjoy!
1. Be the solution.
When mistakes happen, how you communicate sets a critical tone.
"I'm sorry that happened. I don't know how they missed that bug. I also don't know how that other team messed up. And it's a shame that third communication was messed up by that PR person. Yeah, bummer."
"I'm sorry that happened. I should have caught that bug, noticed that error, and I should have reviewed that communication. I'm going to go make certain we don't have any more issues."
People will blame you (or not blame you) regardless of your words. It's foolish to point fingers when it rarely changes who will be at fault in people's minds.
- If they blame you, and you dodge ownership, they'll be annoyed.
- If they don't blame you, and you dodge ownership, no one cares.
- If they don't blame you, and you take ownership, they'll be impressed.
- If they blame you, and you take ownership, they'll be reassured that you admitted it.
Regardless if it's 5% your fault or 95% your fault, make yourself the solution. Own it, and move on with the solution.
2. Measure what you need.
This is a bit of a technical topic, so feel free to skip if you're not technical. Or read it because why not. I remember reading a very interesting article about fashion design recently, and I wore a big purple bathrobe to work more than once. So I feel it's very reasonable to read about topics where you might not be an expert.
When people need to monitor a system, they too often start with what they can measure. Because that's the easy way out.
"We launched this website, and here's a graph of the number of webpage impressions."
Sure, you can graph that, but why do we care? Was that something we needed? Is that graph useful? I mean, it could be interesting, but does it solve our needs?
Don't do things backwards by starting with your capabilities. Always start with the customer. That customer could even be you.
Why do you require metrics?
- "I'm worried about having an outage, and we don't know about it."
- "I'm worried that our pages might load slower over time."
- "I worry that our registration pipeline could break without us knowing it."
Great. Now we have a list of requirements, so you need metrics to monitor those things.
But not just those things. You want the coarsest metrics you can find.
Why coarse metrics?
Because metrics are primarily for identifying that there is something going on, and you should research it. Essentially, "Alert! Pay attention!"
They're not for identifying what went wrong. If your system breaks, you can surely figure out why, as long as you know it's broken.
If you have a website, you can catch 500 errors and report on them. But what about 503 errors? Or 403 errors? Yes, you'd surely want to know about any type of error. One way to approach that is to build 37 metrics and 37 alarms for each type of error. But what if you forget one? It's a poor way to monitor a system.
Instead, the right answer is to build a coarse measurement, one that catches all errors. So, something like two coarse measurements will catch almost everything. That's not bad at all!
- Measure all errors for all customers. You'll identify most problems.
- Measure the rate of customer's successful actions (e.g., purchases). This'll probably help you notice the rest, because if there's a big issue, that rate would drop.
Graphing metrics is also a valuable tool, as long as three major things are true:
- Have a regular cadence for reviewing them (e.g., some daily, some weekly).
- Every time you have an issue you didn't see in your graphs, figure out why, and if your graphs need to change.
- Don't have so many graphs that you don't look at them all.
It's astonishing how often business with brilliant engineers like AWS still found events by a human looking at a graph and saying, "Hey! Is that a bump at 3am every day? What's going on?"
Alarming on metrics is valuable, as long as:
- You don't have so many false alarms that you begin to ignore them.
- You don't miss events because your thresholds are too lax.
3. Nail the basics before going bigger.
If you're interested in career growth, it begins by doing your current job extremely well.
People frequently ask for stretch assignments when they're dropping things on the floor. It rubs their managers and co-workers the wrong way. As in, "Holy smokes, this is more evidence that you have poor judgement."
Never ask for a promotion if you're receiving constructive (negative) feedback on your current work. It's a bad message send because it suggests you don't have your priorities straight.
4. The internet wasn't around when I was a kid.
My 9-year-old daughter said it's interesting how the internet and devices weren't around when I was a kid.
She said they're everywhere now, and so important. I agreed.
She wondered what would be the thing which would be everywhere and important when she gets older. I said that was a good question.
She suggested it could be a pill which lets you fly a couple of feet off the ground. It'd allow you to get places faster.
I hope one of you is working on this pill.
5. Learning about customer needs.
As people say, customers don't always know what they want. You can certainly create something customers don't realize they want. For example, the Ember mug. I had no idea I wanted a mug which stayed warm using battery power. But I realized that I did want it, once I knew it existed.
But even if they don't know what they want at first, customers still have agency over their time and money.
That means there are two major challenges with identifying customer need.
One - Customers need to eventually believe they have this need, or they won't use your thing or part with their cash.
While customers may not know what they want, you have to have a clear and simple path to explain why they will want it.
Here are three potential results from learning customer needs.
- "Here is an email client which stores gigs of email for free." "Awesome! That's something I really wanted!"
- "Here is a voice assistant which can play music and set alarms via voice." "Cool! I didn't know I wanted that, but I do!"
- "Here are some hideous glasses which could take low-quality videos of your surroundings." "Um... no thanks."
Two - Your solution needs to fulfill their needs, or you won't have long-term success.
"Use this marketing tool to increase sales." "Sure, I love increased sales."
But those sales going up or not are what makes this a happy recurring customer, or an unhappy ex-customer.
- If customers already want it, you need to be effective.
- If customers don't already want it: 1) You need to verify that they should want it, 2) That you can sell it to them, and 3) You can be effective.