Digital Marketing · Advertising
Solace
Launching a meditation app to the people who would never call it that — one idea, five channels with jobs, and a retention plan written before the first ad ran.
- Sector
- Health & Wellness
- Disciplines
- Campaign, Media
- Scope
- 10 Weeks
- Deliverables
- 26 Assets
For the 11pm brain
Put the day
down.
Nine minutes, a voice, and a timer. No incense, no journey, no streak to break.
Say what's keeping you up
One question, four answers. No account yet.
Get one session, not a library
Nine minutes, chosen for you, playable offline.
Put the phone face down
The screen dims itself. Nothing to tap again.
Campaign landing page — headline matched to the ad that sent you
Overview
Solace is easy to install and easy to forget. The campaign had to solve the second half.
Solace is a self-initiated concept project — a fictional meditation app built so we could work through the launch problem every subscription app eventually hits: the media is the cheap part, the second session is the expensive part.
We ran it as a full engagement: audience definition, positioning, a creative platform, the ad units themselves, a channel plan with a job per line item, and the lifecycle sequence that catches people after install. Every asset on this page was written and designed for the project.
The Challenge
A category that only talks to people who are already calm.
Wellness advertising has settled into a single template: a dusk-lit landscape, an indigo gradient, a soft promise about your journey. It is beautiful, it is interchangeable, and it is addressed to somebody who has already decided that meditation is for them. That audience is small, expensive, and mostly spoken for.
The second problem is structural. An app install is a low-effort action, which makes it a cheap thing to buy and a misleading thing to celebrate. The moment that actually decides whether Solace has a business is the one that happens the following night, when nobody is being advertised to. If the ad promises transformation and the product delivers nine minutes and a timer, that second night never arrives.
Earn attention from a sceptic
Reach people who would never search for a meditation app, using the specific problem they would describe out loud — not the solution category they would never name.
Make a promise the product can keep
Set the expectation at nine minutes, so the first session lands as delivery rather than disappointment. The ad is the first half of the onboarding.
The category, and the gap in it
Again.
Four category-typical layouts and ours, at the size a person actually sees them. The difference isn't the palette — it's that one of them says something a specific person would recognise as their own Tuesday night.
The Audience
Three people, one product, three completely different first sentences.
We wrote the audience as three people rather than one demographic, because the demographic they share tells you nothing about which sentence will stop them. Each one arrives through a different door and needs a different opening line — but the promise underneath is identical, which is what keeps it a campaign rather than three campaigns.
The 11pm scroller
Late twenties to late thirties. Sleeps badly two or three nights a week and has never once used the word mindfulness about it.
"I'm not stressed. I just can't stop thinking."
Reached via
Short-form video after 22:00, podcast host-reads on late-night listening.
The returner
Has already downloaded two apps in this category and lapsed from both. Doesn't need convincing that it works — needs permission to start again.
"I did four days and then felt bad about it."
Reached via
Retargeting, creator content that admits to lapsing, guilt-free lifecycle copy.
The sceptic
Thinks wellness has become a personality. Allergic to the vocabulary, not to the outcome. Will try something that is described plainly and priced clearly.
"I'm not going on a journey."
Reached via
In-stream video where the product can be shown working, non-wellness creators.
One promise, three doors
Who we deliberately excluded: experienced meditators. They are already served, they cost the most to convert, and writing for them is precisely what has made the whole category sound the way it does.
The Product
The ad promises nine minutes. The product has nine minutes to prove it.
We designed the first four screens alongside the creative, because they are the same argument continued. The onboarding asks the question the ad implied, the session delivers the exact duration the ad quoted, and the progress screen refuses to punish a missed night — which is the promise the returner segment was sold on.
What's keeping
you up?
One answer. Change it any night.
No account needed yet
Good evening,
Nadia.
Tonight
Putting a thought down
Also short
Three minutes, eyes open
3 min
The 3am wake-up
6 min
Sunday night, specifically
11 min
Putting a thought down
Marcus · 9 min
4:07
Breathe out
Screen dims in 10s
Your nights
You showed up
four nights.
Four more than none. Nothing resets.
Noticed
You start earlier on Sundays. Move tonight to 21:15?
Minutes, all time
142
No streaks here. We removed them on purpose.
Onboarding, Today, Session, Progress — the four screens the campaign is accountable to
Colour
Sand and Ink carry the layouts. Sage marks exactly one action per screen. Mist appears only in motion — the breathing ring, the progress arc, the second a session ends. Nothing in the palette glows, because the product is used with the lights already off. Choosing warm sand over the category's midnight indigo also does the media a favour: in a feed at 11pm, every competitor is a dark rectangle on a dark screen.
Typography
Timer numerals are set tabular so the countdown doesn't jitter — a detail nobody notices until they're staring at it in the dark.
The Creative
One idea, six placements, nothing rescaled.
The platform is Put the day down. Every unit names one specific unfinished thought, then offers a fixed, unmystical amount of time to hand it over. The tension changes; the promise never does.
Structurally, each unit holds three things constant — the mark in the same corner, the sand-and-deep palette, and the nine-minute offer — and varies exactly one: the opening line. That discipline is what makes a test readable later. If everything moves at once, a winning ad tells you nothing about why it won.
In-feed, 1:1
solace
Sponsored
You are still writing that reply in your head.
Solace has a nine-minute wind-down for exactly this.
Try free for 14 daysCold prospecting. The line does the targeting — anyone it doesn't describe scrolls past, which is the point.
solace
Sponsored
Nine minutes. Then bed.
Two taps to tonight's session. No streaks, no guilt, no library to get lost in.
Open SolaceTonight
Putting a thought down
9:00
Retargeting for installers who never started a session. Shows the product, not the promise — they've already heard the promise.
Vertical stories & reels, 9:16
11:47pm.
Again.
It isn't stress. It's just the day, still running.
Awareness. Runs after 22:00 only.
"I can't switch my brain off."
Nobody can.
Solace isn't switching anything off. It's nine minutes of somewhere else to put it.
Objection handling for the sceptic. Concedes the point in line two.
Still up?
Bed at midnight either way. This takes nine of the minutes in between.
Late retargeting. Same idea, dayparted to the moment it describes.
In-stream video end frame, 16:9
Meditation with the mysticism removed.
Nine minutes. A voice. A timer. That is the entire app.
Skippable in-stream, cut so the product is visible before the five-second mark. The end frame carries the price, because the sceptic segment leaves when the price is hidden.
What we held constant, and what we refused
Constant across every unit
- The mark, same corner, same size relationship.
- Sand or deep. Never both fighting in one frame.
- The nine-minute number, stated as a number.
- One CTA, one verb.
Ruled out on day one
- Sunsets, mountains, and anyone sitting cross-legged.
- The words journey, ritual, mindful, and unlock.
- Whispered voiceover as a substitute for a script.
- Any claim about sleep quality we could not evidence.
The Funnel
Every stage gets a different sentence and a different number.
A campaign that says the same thing at every stage is really a campaign with one stage. We wrote the message and named the diagnostic signal for each step, so that a flat week can be traced to a specific point rather than blamed on the creative in general.
Awareness to retention
Awareness
"11:47pm. Again."
Reels, TikTok, in-stream video
Consideration
"Nine minutes. A voice. A timer."
Landing page, podcast reads, retargeting
Install
"No account needed to start."
Store listing, search capture
Activation
"First session tonight, not tomorrow."
Onboarding question, first push
Retention
"Skipped last night. That's allowed."
Push and email lifecycle
Awareness
"11:47pm. Again."
Reels, TikTok, in-stream video
Consideration
"Nine minutes. A voice. A timer."
Landing page, podcast reads, retargeting
Install
"No account needed to start."
Store listing, search capture
Activation
"First session tonight, not tomorrow."
Onboarding question, first push
Retention
"Skipped last night. That's allowed."
Push and email lifecycle
Signal, top
Three-second view rate and landing-page click-through, read per creative line rather than per campaign.
Signal, middle
Store-listing conversion rate, split by which creative angle referred the visit.
Signal that decides it
Second session within seven days. Every other number on this page is a leading indicator of that one.
Activation is highlighted because it is where the campaign is actually judged. An install that never becomes a session is a cost, not a customer — and it is the only stage where the fix might be a product change rather than a media one.
Channels, and the job each one is hired for
Short-form paid social
Awareness · 40%
Where people who don't identify as meditators are reachable at scale, sound-on, at the hour the creative describes. Also the cheapest place to run six versions of a line and find out which tension is real.
Online video
Consideration · 20%
The only placement with enough seconds to show a session as a session. For the sceptic, seeing the app do the plain thing it claims is worth more than any line of copy.
App-store & search capture
Harvest · 20%
Demand the campaign creates has to land somewhere. This line harvests intent already in motion — "can't sleep app", "sleep sounds" — and it should be judged as capture, not as growth, or it will flatter itself.
Podcast host-reads
Trust · 12%
Late-night listening puts the ad inside the exact moment it describes, read by a voice the listener already tolerates at 11pm. Scripted to be spoken, not read — the host says what kept them up, then the offer.
Creator partnerships
Credibility · 8%
Deliberately not wellness creators. The pitch lands harder from someone who has never told their audience to breathe before, and whose feed the sceptic hasn't already learned to discount.
Owned lifecycle
Retention · 0% media
Costs nothing in media and carries the entire retention job. Written in week two alongside the ads, not in month three when the churn report arrives.
The split we would open with — and expect to be wrong about by week three. It is a starting hypothesis about where attention is, not a forecast; the first read on second-session rate by source is what would move it.
Lifecycle: the part that decides whether any of it worked
Every message below is sent relative to the bedtime the person gave us during onboarding, not at a global 8pm blast. And none of them imply failure — the returner segment lapsed from two other apps precisely because those apps made a missed night feel like a broken thing.
Push sequence
Solace
21:40Tonight's session is ready. Nine minutes.
You said it's the thoughts. This one is for the thoughts.
Solace
21:40Skipped last night. That's allowed.
Nothing reset. Tonight's is still nine minutes.
Solace
21:15You start earlier on Sundays.
Want tonight's a bit earlier too? Two taps to change it.
Solace
21:40Trial ends Thursday. No surprise charge.
Decide on Thursday, not now. Tonight, just do the nine minutes.
Four pushes in twelve nights. The cap is the strategy: a wind-down app that interrupts you is arguing against itself.
Day 1 · Email
You did one. Here's the one that follows it.
Preview: It's shorter. On purpose.
One is not a fluke.
Last night you did nine minutes. Tonight's is six, because the second one is easier when it asks for less. It's already queued.
Open tonight's sessionDay 9 · Win-back
We noticed you stopped. No lecture inside.
Preview: One question, then we'll leave it.
Was it the voice, the length, or the hour?
Pick one and we'll change it. Pick none and we'll stop emailing — that link is at the bottom and it works in one click.
The win-back email offers the unsubscribe in the body rather than the footer. A wellness brand that makes leaving difficult has lost the argument it's making.
The Result
A campaign that promises exactly what the product does, so the product gets a second night to prove it.
One idea that survives the crop
The platform holds at 1:1, at 9:16, at six seconds, and in thirty characters of push copy.
Channels with actual jobs
Every line item has a stated purpose, which is what makes a bad week diagnosable instead of debatable.
Retention written with the ads
The lifecycle sequence and the creative came out of the same week, in the same voice.
What we'd put in front of real spend first
That naming the thought beats naming the feeling
The assumption we'd validate first: a specific 11pm thought out-performs category-standard calm imagery. Same offer, same targeting, two creative systems, nothing else different.
That the sleep door is worth paying more for
We'd expect the sleep angle to cost more per install than a broad wellness pitch, and be worth it on second-session rate. If it isn't, the positioning is wrong and we'd say so.
That removing streaks retains the returner
The riskiest belief on the page — the whole anti-guilt lifecycle rests on it. It's a straight holdout against a conventional streak mechanic, and it's the one we'd most want to be right about.
Where a results slide would normally go. Solace has never run. No media has been bought, no installs exist, and there is nothing to report — so instead of a wall of invented percentages, this page shows the thinking, the creative, and the measurement plan we'd be held to. On a live account, this section would carry whatever the tests actually returned, including the two of the three above that probably wouldn't survive them.