ia

Currently shipping at Threadline

I help teams ship products that don't need a marketing team to explain them.

PM for 0→1 consumer products. Nine years in, mostly spent in the unglamorous middle — between the customer who can't find the button and the engineer who insists the button is obvious. I make the call about who's right.

11

PRODUCTS LAUNCHED

Three I'd ship again. The rest I'd kill faster.

4.1M

USERS TOUCHED

Across consumer mobile, B2B SaaS, and one regrettable web3 detour.

$148M

ARR INFLUENCED

PM math is fuzzy. I'd defend $90M of that in a board room.

27

FEATURES KILLED

The metric I'm proudest of. See section 06.

Every roadmap I've inherited had three features nobody used and one feature everybody hated. My first quarter is usually a removal exercise.

Once the noise is gone, the product usually tells you what it wants to be.

I.

The customer is never the customer.

What they say they want is downstream of an unspoken job they're trying to get done. Most PMs build the request. The good ones ask one more question.

II.

A roadmap is a memory aid, not a contract.

If a Q3 launch makes more sense in Q4, we move it. If a feature stops mattering, we kill it. The plan is in service of the work — not the other way around.

III.

Taste is a deliverable.

Not the squishy kind — the kind that lets you reject a beautifully-designed feature because it doesn't belong. PMs who can't do this end up shipping museums.

THREADLINE · ONBOARDING REBUILD · Q4 '25

We rewrote onboarding by deleting it.

READ IT IN

~7 min

THE PROBLEM

31% D1 retention.

Threadline's onboarding was 11 screens long, and 69% of users dropped before reaching the product. The team's instinct was to optimize each screen. Mine was to ask why the screens existed at all.

We were collecting data we never used, asking for permissions we didn't need yet, and teaching features users hadn't earned. It read like a contract, not an introduction.

THE BET

Keep three screens. Earn the rest.

Sign-up, one personalization question, and the home screen. Everything else — permissions, account setup, the tour — would be triggered contextually, the moment a user actually needed it.

Engineering pushed back. Marketing pushed harder. I had to make the call that the metric we'd be measured on (D1) outranked the metric they wanted to defend (data completeness at signup).

"

Iris is the kind of PM who'll let an entire feature die in a meeting if she's not convinced. It's annoying right up until you realize she's saving you six weeks.

— Sasha Velez, VP Engineering at Threadline

RESULT

+21pts

D1 retention, week 7

Customer interviews, support tickets, sales call recordings, the Slack channels where engineers complain about the product. The unfiltered stuff. I don't take a stance for the first two weeks.

One page. The problem in plain English, the constraint, the bet I want to make. If I can't say it without jargon, I don't understand it yet. The memo lives — every meeting refers back to it.

First version goes to 1% of users in three weeks, even if it's ugly. The point isn't to be right, the point is to be informed. Iterate on what users actually do, not what they say in a survey.

Especially when it works. Wins teach you less than losses, but only if you write down what you actually believed at the time. I keep a decision journal. Reading it back is humbling.

D1 RETENTION

We cut onboarding from 11 screens to 3. I was told the data team would never approve it. I wrote a memo. They approved it.

USERS, 0→1 LAUNCH

The budgeting product. We hit 1M in 14 months with zero paid acquisition. The plan called for $4M in ads. I argued we didn't need them. We didn't.

SUPPORT TICKETS

Patient intake flow. Read 400 support tickets, found three patterns, fixed two with copy changes. The third needed engineering. Took two weeks.

SIGNUPS FROM REFERRALS

A referral loop that didn't bribe anyone. People shared because the product made them look good. We removed the discount code two months in.

"

She let the room argue for forty minutes before saying anything. By the time she did, we'd talked ourselves out of the bad idea. I think she planned it that way.

"

Iris is the only PM I've worked with who'll say 'I was wrong' in writing, then attach the data. It changed how the whole team treated post-mortems.

"

Worked with her on the patient intake redesign. She wrote a memo so good our hospital partners adopted it as their internal training material. Quietly devastating writer.

I'm open to senior PM roles, fractional engagements with seed→Series B teams, and the occasional advisory call if your roadmap feels haunted. I respond inside 48 hours, even when it's a no.

ELSEWHERE