Values

Every values page reads the same. Integrity. Excellence. Innovation. Words picked because they sound good in a pitch deck. Nobody checked what keeping them would cost.

I sat down to write this page the same way and got three sentences in before I threw it out. Every value I'd picked felt performed the moment I tried to explain why it mattered.

So I started over with a harder rule: nothing goes on this page unless I've actually had to live by it, like killing a feature after building it, sitting through a client conversation I didn't enjoy, or tracking my own confident calls for six months and finding out how often they were wrong. That cut the list by more than half.

Six made it.

Understanding

AI can hand you a plan, a brief, or a first read of a dataset before your next meeting. I use it for all of that, and for a good share of the thinking too. It surfaces options I wouldn't have found alone.

What I don't hand over is deciding whether any of it is right. I've caught myself typing a prompt before I'd worked out what I believed, and watched the answer come back sounding just as confident either way. The answer couldn't tell the difference. I had to.

So when I bring you a recommendation, I've done the understanding myself.

Staying correctable

You've probably reread your own deck three times and missed the typo a colleague spotted in five seconds. Looking harder didn't help. You were reading what you meant, not what was there.

My judgment works the same way, so I stopped trusting how sure I feel. For six months I tracked every confident call I made and checked it later against what happened. Some calls that turned out fine were wrong reasoning that got lucky. Some I doubted most were right for reasons I hadn't seen going in. A good outcome told me the bet paid off. It didn't tell me the bet was sound.

Winning makes this harder. I've felt it on small wins: a client praising a call I got lucky on, a post landing better than the argument deserved. People agree with you a little faster after that. So I keep things around me that can tell me I'm wrong: a second opinion, a client's confusion, a number that refuses to move.

Owning the outcome

You've probably sat in the meeting where rankings are up, traffic is up, and revenue hasn't moved. Everyone at the table can show their part worked.

I've been the one whose part worked. I've built features that did exactly what the spec said and moved nothing that mattered. Finishing felt like solving until I checked.

"That's not my department" doesn't change what the customer gets. So I ask whether something changed for the person using it before I ask whether it shipped. Then I follow the problem past my role, my title, my discipline, until I find the part that's actually broken.

Straight answers

"It's handled" is the sentence I reach for when I want a client to stop worrying. It's also the sentence I trust least when someone says it to me. It tells you how the speaker feels, not what happened or what almost went wrong.

Calm is easy to give you. Understanding takes longer, and sometimes it means telling you about the thing that nearly broke.

So I try to tell you what's true before you have to ask.

Honest priorities

Ask someone what they're prioritizing this quarter and you'll get a clean answer. Then look at where their money and their hours went. The two rarely match. That gap usually isn't dishonesty. Describing a strategy is easier than living inside what it costs.

I hold myself to the same test I'd apply to a client: what my own calendar and spending actually show. I won't pretend that's a settled score. It's a standard I check myself against, not a receipt I can hand you.

Seeing for myself

You can add a metric to the dashboard every quarter and still be surprised by the next lost deal. Tracking has holes. Every metric answers a question you already knew how to ask.

I've watched a good decision get killed a quarter early because the numbers hadn't moved yet. Nobody could tell whether it had failed or just hadn't had time to work. The dashboard couldn't say. Someone had to go look.

Surveys have the same limit. They reach what someone's willing to say out loud, and most of what moves a decision never gets said. So I'd rather go find out what's actually happening than trust a number that can't push back, even though it's the slower way to be right.