Thomasbcn

Independent mind working on mobile growth. Not Ex-GAFAM.

[Thomasbcn @Thomasbcn]

Marketing 101: I've had this awesome "features vs benefits" pix from @UserOnboard pinned since 2017.

Here's my additional 2cents:

[Eric Seufert @eric_seufert]

OpenAI introduces conversion optimization, app install attribution, and advanced matching

“This new set of features represents a substantial update to OpenAI’s self-serve advertising platform, and it materially closes the functionality gap between OpenAI’s offering and those of other, leading scaled platforms — but especially Meta’s.

To my mind, the three most meaningful new features enumerated here are oCPC conversion optimization, mobile measurement partner integrations, and automated advanced matching.”

https://t.co/Ly3kIw6lY2

[Thomasbcn @Thomasbcn]

@ivesparrowai No idea why their reps are so condescending.

The opposite of Meta/Google/Apple who are very nice, yet useless...

[Simon Taylor @sytaylor]

Wow. Revolut just confirmed its $115bn valuation according to WSJ 🤯.

Staggering. I'm shook.

A company that launched as a travel money card in 2015 is now worth more than Barclays. More than Deutsche Bank. It's the first private company Europe has ever produced worth over $100bn, and among private fintechs anywhere on earth, only Stripe ($159bn) sits above it.

Its also the 5th largest private company in the world, only Anthropic, OpenAI, Stripe and Databricks beat it.

The number underneath it is even wilder.

Employees are cashing out at $2,017 a share. There's about to be one HECK of a Revolut Mafia starting new companies.

What earned it:

- A full UK banking licence in March, five years after first applying
- A US national bank charter application filed (with good odds)
- 65m+ customers across 45+ countries, with India, Mexico and Colombia in motion
- $6bn of revenue and $2.3bn of pre-tax profit in 2025
- New licences in UAE, Australia, Mexico and Brazil

This is the first truly global bank.

Now do the maths on the price. $115bn on $2.3bn of pre-tax profit is roughly 50x. That's a software multiple attached to a licensed, deposit-taking bank. Investors are pricing Revolut as a tech company that happens to hold your money.

And look at who's paying it. Coatue, Fidelity, T. Rowe Price, a16z. Nvidia's venture arm joined the cap table in November.

Think about that. In 2026 the marginal growth dollar has one obvious home: AI. This one went to a bank. Even Nvidia, the company the entire AI boom runs through, wanted a piece of a bank (albeit, one that created a custom foundation model that runs on Nvidia hardware).

Because Revolut is that rare thing, a bank compounding like a tech company.

Often, my American friends underrate this company because it hasn't landed on their shores yet. The US charter is the swing variable. If it's granted, Revolut gets a shot at the largest banking market on earth, and $115bn will look like the early price.

Never bet against Revolut.

[Eric Seufert @eric_seufert]

An alternate information space seems to have developed around the chatbot advertising market that has led people to believe that OpenAI's advertising product is an abject failure, ads in chatbots are fundamentally nonviable anyway, and the entire category will never generate a meaningful amount of revenue.

Meanwhile, in the primary information space, we know that Alphabet saw Search queries reach an all-time high in Q1 while Search revenue grew by 19% and ads in AI Overviews monetize at parity with those in traditional Search. The company is expanding advertising in its AI-generated and chatbot surfaces cautiously but deliberately: it introduced formats like Conversational Discovery ads and Highlighted Answers in AI Mode in May of this year. The latest usage numbers shared by Google are 2.5BN MAU for AI Overviews and 1BN MAU for AI Mode. AI Overviews is roughly two years old; AI Mode is roughly one year old.

Further, we know that OpenAI reached $100MM in annualized advertising revenue a mere six weeks after introducing advertising with what was essentially a manually operated product. Since then, OpenAI has executed at breakneck speed to introduce new features: CAPI and pixel integrations in May and Custom Audiences last month. ChatGPT now serves more than 1BN MAU and was already serving more than 900MM WAU in March. ChatGPT launched at the end of 2022.

OpenAI's advertising revenue goal for 2030 is incredibly ambitious; I have no affiliation with OpenAI, and I don't have any quantitative indication of whether they'll meet that goal. But to conclude that OpenAI's advertising efforts are somehow failing now, or that the market won't evolve into something meaningful, feels to me totally detached from the primary information space.

[Thomasbcn @Thomasbcn]

@ivesparrowai @filippkowalski I did see 1 case with a gap in result, and heard about various others. Kinda weird to me. Potentially the mmp was poorly implemented/set up there

[Thomasbcn @Thomasbcn]

@ivesparrowai @filippkowalski I dont think there is a meaningful gap in performance for the one feature they compete on, but Appstack is better dovumented/easier, potentially cheaper. I’m looking fwd the advanced mode that enables each to set their FP sctrictness/lax, something MMP don’t provide nor disclose

[Thomasbcn @Thomasbcn]

@ivesparrowai @filippkowalski I'm affiliated with nobody, but I have been a customer of all above. What do you want to know about it?

[Thomasbcn @Thomasbcn]

@david_bonilla Prefiero que me dejes colgado que me contesté tu bot

[Thomasbcn @Thomasbcn]

I call the COO "VP of shit"

I would likely suck at it, being after seeing various average and a couple incredible individuals at it, I believe the role is often more critical than any other C

Kudos to the few who can handle the infinite stream of random issues

[Blake Anderson @blakeandersonw]

Hiring a Head of Operations

Company is at $4m ARR growing 20% month over month.

No skills or experience required. Only requirement is willingness to work hard, learn, and drop your ego.

Link below.

[Thomasbcn @Thomasbcn]

@ivesparrowai I’m actually in favor of brand defense, for various reasons.

Blending it with the rest?
Deceit, if not straight fraud.
Clean split is the only way. When the platform doesnt allow it for questionnable motives, testing what it looks without is an absolute must.

[Thomasbcn @Thomasbcn]

@ivesparrowai Yes, and brand can be negated there too
(Took almost a decade for that feature and its burried very deep)

[Thomasbcn @Thomasbcn]

Where can I see my own likes timeline on the iOS app? Not where it used to be

[Vahe Baghdasaryan @vahebagdasar]

We’re hosting two CEO dinners in Paris tomorrow and Thursday, with some amazing app founders joining us.

We’re technically full, but if you really think you should be in the room, send us a DM. We might be able to make it work.

[Thomasbcn @Thomasbcn]

@ivesparrowai Great for you, keep grinding at it if so!

Have you ever tested negating brand terms there?

[Thomasbcn @Thomasbcn]

@eric_seufert Meta, Applovin, Google, Unity et al to benefit most those “savings” for sure

[Hannah Parvaz @HannahParvaz]

You don't need more channels. You need 1 working properly.

[Thomasbcn @Thomasbcn]

@MichalLangmajer I started collecting public figures, then I stopped as

1. Most were just brag
2. Dubious methologies of attribution
3. Variance is extreme for this
4. Averaged benchmarks are useless

A more interesting question is what elements drive higher IPM

[Thomasbcn @Thomasbcn]

@mcflyDev I believe so. If you can without any delay, better, but that removes the possibility to filter out “instant cancellers”

[Thomasbcn @Thomasbcn]

Signal engineering: every hour matters, even within day0

Timo Dechau confirming here what I'm insisting on: time to send conversion back (signal) matters, a lot. The sooner the better.

Meta wasn't that explicit before, but experiments clearly show a steep decay.