Technology tells you what it can do. Culture tells you why it matters.

When Effect House came to France, the story of AR had already been written by the tech giants who got there first. It was either a complex utility for engineers, or a three-second filter for a selfie. It had virality. It had utility. It had no prestige.

Meanwhile a generation of young artists was watching, and wanted none of it. They did not need another digital toy. They were looking for a canvas: a medium that respected their ambition.

So the problem was never distribution. It was cultural:

How do you elevate AR from a technical feature into a cultural medium?

Early research came back with a low NPS, which told us something was wrong but not what. In-depth interviews did: the people who mattered were young designers, and what they wanted was not a better tool. It was exposure and a way to earn from the work.

That reframed the product. Effect House existed to solve three problems at once: not enough effects to meet user demand, brands wanting their own branded effects with no path to get them, and designers with nowhere to monetise what they made.

The original message described the product. The new one described what the product was for.

New technologies are adopted through expression, not understanding.

People do not engage with a tool by learning what it does. They engage by using it to say something about who they are.

So the question stopped being how to teach AR, and became: what would make a young design student pick it up not as a software exercise, but as an artistic statement?

The answer wasn't a tutorial. It was a stage.

A stage needs three things: a reason to create, the right audience, and a space that gives the work meaning.

The reason

A creative contest

An open contest inviting creators to reinterpret the heritage of the Grand Palais in AR, turning a technical exercise into a creative statement.

The audience

Target the right audience

Off existing AR creators, tech communities and mature designers. Onto art and design schools, aspiring creators and young designers.

The space

Inside Grand Palais

Selected works exhibited inside the Grand Palais, placing AR where it could be seen, shared and taken seriously. Creators didn't just show work. They met, and became a scene.

They didn't talk about features. They talked about culture.

What started as a product launch became a cultural moment. The French press covered it as an event about technology and culture rather than a feature release, and one of the contest prizes was awarded for the effect most beneficial to culture. Three thousand people came to the Grand Palais, at an acquisition cost well below what paid channels would have charged for the same attention.

The programme also seeded a creator ecosystem that kept running after it closed: over a thousand new effects a month, four in five of them originating here, and a twenty percent increase in monetisation opportunities, which was the thing the designers had said they wanted in the first place.

Creator effects made for the contest and after it, unprompted and unpaid.

TikTok wanted to hold its position as the music platform of choice, and the way to get there was more music content being made on the platform. But engagement with music content was low on both counts that mattered: how long people stayed, and how much they made.

At the same time the label partners had their own problem. Artist growth was hard, and the formats available to them were not solving it.

So the question was not how to promote music better. It was: where is the product-led growth opportunity for music on TikTok?

Music on TikTok has two customers, not one: the artists who make the content and the users who watch and remix it. They want different things from the same moment, so I ran user research and built a persona for each before touching the solution.

Then I mapped the existing journey against those personas. The map showed something specific: users' emotions dipped during music livestreams. Interest fell exactly where engagement was supposed to be highest.

We had already tried lifting engagement at the other touchpoints, repeatedly, without success. That pattern is usually the signal that the problem is not in the execution.

The problem was the format, not the promotion.

Livestreaming is time-sensitive and crowded. It asks everyone to show up at the same moment, it competes with everything else happening at that moment, and when it ends there is nothing left to return to. For an artist trying to build a fanbase over time, that is the wrong container.

So rather than promote the format harder, I designed a new one.

A landing-page driven music hub, built around two properties the livestream did not have.

Reusable

Not time-sensitive

Landing-page based rather than live, so it stays up, stays findable, and can be templated for the next artist instead of rebuilt.

Engaging

A reason to engage

Missions that turn watching into making, with points and a lucky draw that bring people back rather than letting them bounce after one track. Making was the behaviour the brief existed to move.

A campaign ends. A format keeps running.

Facing rising churn on a $50M ARR product, the company needed a new pricing plan to protect retention. By cross-referencing user interviews with actual usage data, I found that daily users were not always the buyers, and the loudest requests were not the real problems. Cutting through the noise, I identified the single core feature users returned to weekly, paid or not, and anchored our new pricing model directly on it.

200

Quantitative responses analysed

20

User interviews

Two major churn drivers were product flaws that no amount of marketing could mask. I partnered with engineering to overhaul our matching algorithm and redesign the onboarding flow, letting users experience the product’s core value before demanding heavy setup. This tripled algorithm accuracy and boosted onboarding completion by 24%. We deployed these product fixes first, rather than rushing into the pricing launch.

The old access-based pricing was broken: small accounts overpaid for unused capacity, while large ones extracted massive value at a flat rate. I replaced this with a usage-based model that scales naturally alongside the customer’s own growth. I also redesigned the packaging into six distinct “Jobs to be Done” tiers. By keeping the core value in the entry plan and mapping a crystal-clear upgrade path, moving to a higher tier became a frictionless, logical progression rather than a penalty.

Free-to-paid conversion. Not sign-ups, not reach, not a launch-week spike. One goal settles every argument that comes after it, and it decides which activities are worth booking at all.

Each one carried a conversion target of its own. Target divided by the channel’s historical rate gives the audience that channel has to reach; if the reachable pool is smaller than that, the plan is already broken, and it is better to know in week one than in the retrospective.

Seven workstreams, six functions, one launch date. Both of these were agreed before the first asset was briefed, which is the only reason the date held.

Freemium to paid conversion and upsell carried the ARR lift. The churn and onboarding numbers are what made that possible: the product work came first so that the pricing launch had something solid to stand on.

One headline, one subhead, one button, written for a reader who did not exist: the average of three very different people. A page like that is not badly written. It is written for a composite, and a composite never signs up.

Before. A mission statement: what the company wants to do to the industry.

After. A product promise: what you can do, and the button that does it.

Three pages, three different first verbs. Everyday users are told they can swap, liquidity providers that they can earn, developers that they can build. The same protocol underneath, three different reasons to care, and three separate pages rather than three paragraphs on one.

Positioning that stays in a deck is a point of view. Positioning that ships is a decision, because a live page has to choose who it is talking to.

Everyday users · Swap

Liquidity providers · Earn

Developers · Build