Filip Balenko · {DATE}
FILIP BALENKO · FINANCE & STRATEGY APPLICATION

A finance brief
on Phia,
before applying.

I put this together before submitting my application. The numbers come from Phia’s public announcements; the model, the brief, and everything else here are mine. Felt like a better way to show what I’d do in the role than another one-page resume.

Role Finance & Strategy Associate, Phia Filed May 2026 Email filippbalenko@gmail.com LinkedIn linkedin.com/in/filipbalenko
§ 01 · WHAT’S PUBLIC

Phia, as of this morning.

Sources: company announcements and press
Series A close — January 27, 2026
$43M
Total raised
$8M seed, then $35M Series A
$185M
Series A valuation
Roughly 5x the seed price
1.0M+
Users in 10 months
Almost entirely organic
7,200
Brand partners
350M+ products in catalog
+13%
Higher conversion for partner brands
+30%
Lift in new-customer acquisition
+15%
Higher average order value
−50%
Return rates cut by more than half
§ 02 · MY READ

Phia is three businesses stacked.

From the outside, Phia looks like a shopping app. Underneath, it’s really three businesses running in parallel. There’s the demand side: Phoebe, Sophia, 2M followers, 430M views, basically zero paid marketing. There’s the agent itself, which is what everyone writes about. And there’s the new B2B layer (partner dashboards for brands) that the Series A is funding.

The B2B layer is the part I’d watch most closely. Affiliate is a thin-margin business that depends on the consumer side staying hot. B2B is contracted ARR with software margins. Most of the value created over the next 18 months probably comes from the mix shifting between those two, and you can’t see that in a top-line revenue chart. You need them split out from day one.

01.

Take rate trajectory

Affiliate take rates run anywhere from 3 to 15% depending on category and contract. B2B SaaS clears 80%+ gross margin. A blended top-line revenue chart hides which side is doing the work. I’d model the two separately from day one and watch the mix every month.

02.

Partner concentration

7,200 partners is a great headline. The number I’d actually want is what share of revenue comes from the top 50. Concentration risk is invisible until a big partner churns, and by then it’s usually too late to do anything about it.

03.

Founder-led CAC

430M views and 2M followers got Phia to a million users without meaningful paid spend. That’s incredible and it’s also not permanent. The day Phoebe and Sophia post less, or TikTok’s algorithm shifts, paid acquisition has to fill the gap and the real blended CAC shows up for the first time.

04.

Whether the conversion edge holds

The 13% conversion lift is what brands pay for. It works because today’s users came specifically to use a shopping agent, so intent is high. Once the audience broadens, that intent quality drops and the conversion edge probably compresses. The model should bake that in instead of pretending it won’t happen.

05.

ML hiring against runway

The Series A press release specifically called out senior ML hires. Senior ML talent right now is $300–500K all-in. Twenty of them adds $7–10M of annual burn from that team alone, and the runway math is sensitive to how fast offers close. Worth pulling out as its own line in the model, not bundled into G&A.

06.

Everyone else is also building this

Google, PayPal, OpenAI, Perplexity, Klarna are all moving on agentic commerce. Phia’s edge is the founder-driven consumer pull on one side and the brand data flywheel on the other. Both have to keep running. If either stalls, the moat thins fast, so both should be visible on the weekly KPI report and not buried.

§ 03 · UNIT ECONOMICS

The model, with sliders.

Seven inputs drive everything else: monthly revenue, ARR, gross profit, burn, runway, and what the company would be worth at 8x or 12x ARR. The Bear, Base, and Bull buttons snap to a set of assumptions; you can also move any slider on its own.

Inputs
Monthly active users (millions) 2.0M
Conversion to purchase 6.0%
Average order value $90
Affiliate take rate 8.0%
Partners on B2B plan 5%
B2B ARR per partner ($K) $12K
Monthly opex ($M) $3.0M
Outputs · Monthly & Annual
Volume
Monthly transactions
Monthly GMV
Revenue
Affiliate revenue / mo
B2B revenue / mo
Total revenue / mo
Implied ARR
Economics
Gross profit / mo (blended 75%)
Net burn / mo
Runway (months) $28M assumed cash
Valuation lens
Implied EV @ 8× ARR
Implied EV @ 12× ARR

Built from public disclosures. Cash assumed at $28M, which is the Series A close less roughly six months of pre-A burn. Affiliate gross margin around 95%, B2B SaaS gross margin around 80%, blended at 75% after payment processing. Not a forecast of Phia’s actual numbers.

§ 04 · KPI WATCHLIST

What I’d put on the weekly dashboard.

Six metrics. Each one matters more than a more obvious one right next to it, which is why I’d put them in front of Phoebe and Sophia every week.

01 / WEEKLY

Revenue per active user

Take GMV per MAU and multiply by take rate. This one number tells you whether more users are paying off, or whether you’re just adding people who never transact. User-growth charts can look great while RPU collapses, which means the headline can be misleading you for months.

Monetization quality
02 / WEEKLY

Share of revenue from the top 10% of partners

Concentration creeps up quietly. Then one big partner churns and you find out you had a concentration problem the whole time. I’d want an automatic alert the day this crosses 50%.

Concentration risk
03 / WEEKLY

Organic vs. paid acquisition mix

What share of new users is coming from Phoebe and Sophia’s organic content versus any paid channel. If organic falls below 70%, the blended CAC story changes and the model probably needs an overhaul, not a tweak.

CAC defensibility
04 / MONTHLY

Cohort retention at month 3

Of users who bought something in month 1, the share still buying in month 3. The cleanest signal for whether Phia is a real shopping habit or a fun thing people try once. Habit businesses are worth a very different multiple than discovery moments.

Product engagement
05 / MONTHLY

B2B pipeline coverage

Total qualified B2B pipeline divided by trailing quarter B2B revenue. Under 3x is yellow. Under 2x is red. Most of the Series A story comes back to this ratio one way or another.

B2B engine health
06 / MONTHLY

Personnel burn as a share of total burn

ML hiring is the biggest single swing factor in runway. If personnel burn is climbing slower than revenue, fine. If it isn’t, the board conversation needs to happen one quarter earlier than people usually want to have it.

Runway discipline
§ 05 · SAMPLE MONTHLY BRIEF

What I’d send Phoebe and Sophia.

If I’d had the role for a month, this is roughly what would land in their inbox at close. Same template every month: one page, four parts, no decoration.

May 2026 — Finance & Strategy Brief
To: Phoebe, Sophia · From: Filip
Confidential
Distribution: 2
Drafted on close
1 / Headline

Monthly monetized GMV passed plan by mid-month, mostly from the contemporary apparel cohort. Revenue per MAU was up for the third month in a row, but the gain is sitting almost entirely in the top three verticals. Net burn came in about 6% under plan because senior ML hiring is running slower than we modeled, which adds roughly 0.4 months back to runway.

2 / What changed

The B2B dashboard pilot signed partners five and six. Committed B2B ARR is now at a level where I think it starts to matter against affiliate revenue rather than sitting in the rounding. Organic acquisition was 78% of new users this month, still dominant, but it’s the second month in a row it’s dropped. This is the leading indicator we agreed to watch. Conversion lift for partner brands held at 13–14%, so I don’t yet see the compression we were worried about, but the data on the newest cohorts is thin and I wouldn’t read too much into one month.

3 / What I’m watching next month

Three things. First, whether B2B contracted ARR gets above 8% of total revenue. That’s the point where I’d want to split it out on the board chart instead of bundling it with affiliate. Second, whether the organic share keeps falling. One more month of decline and we should probably run a controlled paid test rather than wait until we’re forced to do it under pressure. Third, ML close rate. We’ve offered five and closed two. If that pace holds, the runway slips back to plan within a quarter.

4 / Asks

Two. One: I need a decision on partner-tier pricing for the B2B dashboards by Friday. I’ve built three scenarios and have a recommendation. Two: I’ll send the revised cohort retention model on Wednesday and would love 20 minutes to walk through the assumptions with you before the next investor update goes out.

§ 06 · FIRST 90 DAYS

The first ninety days, week by week.

Specific deliverables, in order. Some are unsexy.

30
Days · Read, listen, rebuild
  • Week 1Read everything. Current model, board decks, last few investor updates. Sit in on partner calls. Don’t ship anything yet.
  • Week 2Rebuild the current month’s model from raw data, line by line, no copy-paste. The goal is to surface every assumption that’s currently implicit.
  • Week 3First version of the weekly KPI dashboard goes to Phoebe and Sophia, with one sentence next to each metric explaining why it’s there.
  • Week 4Pull the first proper cohort retention dataset and share findings. Probably surface three data gaps that need fixing.
60
Days · Build the rails
  • Week 5Stand up partner concentration tracking with weekly alerts when the top decile creeps past a threshold.
  • Week 6Draft a B2B vs. affiliate split P&L for the next board meeting. First version of the chart we’ll probably use for the next year.
  • Week 7Automate the monthly investor update draft. Goal is to get from raw data to first draft in 90 minutes.
  • Week 8Rebuild the runway model with proper ML hiring-pace scenarios baked in.
90
Days · Own the cycle
  • Week 9Own the monthly close and the briefing to Phoebe and Sophia from start to finish. They should be reading a finished doc, not opening a folder of CSVs.
  • Week 10Write the Q1 strategic memo on where the B2B mix is going. Three scenarios, a recommendation, decision.
  • Week 11Build a simple internal tool that lets the team query the FP&A workbook in plain English so they stop pinging me for every number.
  • Week 12Run the first end-to-end quarterly investor update. Phoebe and Sophia review for tone, not numbers.
§ 07 · A note from me

Two companies later, I’d rather do this work
inside someone else’s.

I co-founded my first company, Devola, at 19. Ran finance through 30+ people across the US and Europe and one year of 225% revenue growth. We closed it down in May of last year.

Extra Business Funding came next. I raised $2M as the sole founder, built the underwriting and reporting infrastructure myself, funded 35+ deals, and wound the firm down this April when the economics weren’t where I needed them to be.

Both runs taught me the same thing: I like the finance and strategy craft a lot more than I like running the company. Phia is the kind of place I’d want to do that work. Consumer AI, two founders driving from the front, a business that’s already past zero-to-one and into the messy compounding part. If any of this resonates, please reach out.

— Filip