Wren Hollis is not a real person. Well Made Plays invented this artist so we can show a complete deliverable at full size, with nothing blurred out, without putting a real client's career on the internet. The structure, the depth and the level of detail are exactly what you would receive. The name, the songs, the venues and the numbers are made up.

Well Made Plays · Meta to Spotify Ad Setup

One flight, one event worth optimising for

Wren Hollis · County Line · release 9 October 2026

Not a creative or content document. What the ads say is out of scope; what they optimise for is the whole of it.

01Run cards

One per flight, and the only page you need open at the Ads Manager keyboard. Conclusions only; the reasoning is further down and you do not need it while you are typing.

Flight A · pre-release25 Sep – 8 Oct 2026 · 14 days
Objective
Sales
Optimisation event
Pre-save custom conversion
surface: smart-link page, pixel + CAPI
Budget
$12/day, ad-set level
Geo
United States only
Attribution
Shortest click window the UI offers
never view-through
Creative slots
6, rotating
Learning exit
~50 events / 7 days
expect 24 to 48/day
EMQ checkpoint
Day 3 and day 7, target ≥ 7
Kill
Cost per pre-save above $0.50 at day 7 [ASSUMED]
Scale
Below $0.25 for 3 days, raise 20% per step
Flight B · release and after9 – 23 Oct 2026 · 15 days
Objective
Sales
Optimisation event
On-page save or follow custom conversion
pre-save ceases to exist on release day
Budget
$12/day, ad-set level
Geo
United States only
Attribution
As Flight A
Creative slots
6, three carried from Flight A
Audiences
Prospecting broad, plus retargeting fed by Flight A's pool
EMQ checkpoint
Day 3 and day 7, target ≥ 7
Kill
Cost per save above $1.00 at day 7 [ASSUMED]
a save is a deeper event than a click; the cited band is for clicks
Scale
Not this cycle. Bank the pool instead.

02Conditions precedent

Three gates, each with how it is checked. Event Match Quality is deliberately not one of them: a fresh setup cannot show an EMQ score before it has spent anything, so using it as a launch gate would block a launch that is fine.

The go-live date is conditional on all three boxes.

03Problem and outcome

Problem

Every link Wren has ever posted goes straight to Spotify, so Meta has never been told what a good outcome looks like. There is no pixel, no custom conversion, and no audience that has ever been built. A campaign run today would optimise for the cheapest tap available, which is exactly the traffic worth nothing.

Outcome

One optimisable event on a page we control, fired by someone taking a library action rather than a click. Success for this cycle is 400 or more pre-saves and saves combined, at or under $0.40 per event, and a retargeting pool large enough that the January single starts warm instead of cold.

04The system, tiered by solvency

Tier is assigned by arithmetic, not by ambition. The gate is a daily budget at or above the ten-dollar floor, and that budget divided by the expected cost per event landing near or above seven and a half events a day.

Tier 0Groundwork only. No conversion spend until the pixel and the link exist.Completed 4 Sep
Tier 1Conversion stack. One campaign, one ad set, six creative slots, broad targeting, retargeting layer in Flight B.Both flights sit here
Tier 2Engagement first, for budgets too thin to sustain conversion learning.Not needed
Tier 3Scaled Advantage+ ceiling. Wants roughly a hundred a day to learn.Far above this envelope

The solvency check, shown

$12 a day against an expected $0.25 to $0.50 per pre-save gives 24 to 48 events a day, which clears the seven-and-a-half threshold several times over and puts the ad set past the fifty-events-per-seven-days learning exit inside the first week. Tier 1 is therefore the correct assignment and Tier 2 would be leaving signal on the table. No collision: nothing else is live in the account, so no mid-flight collision handling applies this cycle.

05Payment and bid mechanics

SettingSpecWhy
ObjectiveSalesThe only objective that can carry a custom conversion. Traffic and Awareness are never used here; they optimise for the tap, which is the thing that is already cheap and already worthless.
BiddingLowest cost, automaticManual bid caps at this budget starve delivery before they save money.
AttributionShortest click window offered; never view-throughThe seven-day-view and twenty-eight-day windows were removed in January 2026. A view-through window at this budget would credit the campaign for listens it did not cause.
Daily floor$10/day, running at $12Below ten dollars a day, delivery is too thin to learn stably even when the event arithmetic looks solvent.
PlacementsAdvantage+ placements, automaticRestricting placements at this spend reduces the auction pool without improving anything measurable.
GeoUnited States onlyDefault is tier-one countries. Narrowed on artist data: 81% of existing listening is Texas, and there is no reason to pay for auctions in markets with no draw and no tour.
Creative rotation6 slots minimumCreative is now effectively the targeting layer. Fewer than five or six slots and the system has nothing to choose between.
EMQChecked day 3 and day 7, target ≥ 7In flight, not at launch. Below seven, fix the match keys being passed rather than the ads.

Benchmark bands used for planning, and the one place they are stretched

The cited band below is for a click through to Spotify. Both flights optimise for something deeper: a pre-save on the landing page, then a save or follow after release. A deeper event costs more per event than a click, always, and no published band for either one held up when we went looking. So the pre-save threshold carries the click band forward unchanged, which is deliberately strict, and the post-release save threshold is set at double it as a labelled assumption. Read both as a line to beat, not as proof of failure: after Flight A closes you will have your own cost per event, and that number replaces these.

Cost per Spotify click-through: $0.25 to $0.50 is the acceptable band and $0.15 to $0.25 is a well-run campaign. Planning here assumes the upper end. On cost per thousand impressions the sources conflict, roughly $8 to $15 against about $20 in the US, so this plan uses the higher figure until that is resolved. Planning with the pessimistic number is why the budget above is solvent rather than nominally solvent.

06Tracker edits implied

Wren has a dashboard but no ads table yet, so this creates one rather than editing one.

WhereChange
DashboardAdd a flight row to the release panel: flight name, dates, spend to date, events, cost per event.
Ads · FlightsNew table. One row per flight, predictions pre-filled from the bands above, actuals columns left blank until day seven after the flight ends.
Weekly action listTwo recurring checks during each flight: day three and day seven.

07Scope

In scope

Pixel and CAPI setup, smart-link build, custom conversion registration and AEM ranking, campaign and ad-set structure, audiences and retargeting, the settings above, the two run cards, the flight log, and the day-three and day-seven checks.

Out of scope

Ad creative and content. Other paid lines, including Spotify Marquee, Ad Studio, playlist-pitch marketplaces and boosted posts elsewhere. The budget envelope itself, which is the artist's to set. Pay-per-stream services, which are declined rather than merely excluded.

08Dependencies and assumptions

ItemStateBlocking?
Business Manager account[ASSUMED] exists under the artist's own loginYes. Confirm before any spend; an account created by someone else is a problem to unwind later.
Pixel installedNot installed. Part of this build.Yes, gate one.
Domain verified[ASSUMED] not verifiedYes for AEM ranking. Half a day if the domain registrar login is to hand.
Smart-link toolTo be chosen. One known tool carries both a Meta and a Spotify pixel; the rest are Meta only.Yes. Pricing is as advertised at time of writing and is not a quote.
Spotify CAPI second railAvailable only if the dual-surface tool is chosen. Optional this cycle.No.
Lookalike thresholdNot reachable this cycle at this pool size. Revisit after the January single.No.

09Acceptance criteria

#CriterionBy
1All three conditions precedent checked and screenshotted22 Sep
2Six creative slots loaded, campaign built, not yet published23 Sep
3Flight A live25 Sep
4EMQ at or above 7, or match keys fixed28 Sep and 2 Oct
5Flight A closed, pool saved as a Custom Audience8 Oct
6Flight B live on release day with the retargeting layer attached9 Oct
7Actuals entered against predictions in the flight log30 Oct

10Next steps

WhenWhatWho
By 10 SepConfirm the Business Manager account is under the artist's own login [blocking]Artist
By 12 SepVerify the domain [blocking]Artist, with us on a call
By 15 SepChoose and set up the smart-link tool, install pixel and CAPIUs
By 22 SepWalk the three gates together, live in Events ManagerBoth
25 SepPublish Flight AArtist
30 OctEnter actuals against predictions in the flight log. This is a named job with a date, not a hope that someone remembers.Artist
Early NovGrade the flight together and decide whether January runs the same shapeBoth

11Sources and verification ledger

Every number above traces to a source line, or is labelled as an assumption. Nothing is cited from memory.

FactUsed forVerified this runMethod
F-MIN-DAILYThe ten-dollar daily floorYesRe-checked against current practitioner guidance
F-META-LEARNINGFifty events per seven days to exit learningYesUnexpired, structural
F-META-ATTRIBUTIONRemoved windows, shortest-click settingYesFetch, Meta documentation
F-META-OBJECTIVESSales objective, never TrafficYesChecked against the current Ads Manager labels
F-EMQ-TARGETEMQ at or above 7, day 3 and day 7YesUnexpired
F-BENCH-CLICKCost-per-click-through bandsYesSearch, re-run this cycle
F-BENCH-CPMCPM planning figureYes, conflict unresolvedTwo sources disagree; planned with the higher
F-BENCH-SAVECost-per-save kill threshold, Flight BNo published band foundAssumption, labelled in the run card: click band doubled, because a save is a deeper event than a click
F-META-ANDROMEDABroad targeting, six creative slotsYesSearch
F-TOOL-*Smart-link tool comparison and pricingYesFetch, vendor pages. Pricing as advertised, not a quote.

Flight log handed over with both flights pre-filled: predictions from the bands above, actuals blank, and a dated row naming who enters them and when.

Well Made Plays

This is a real deliverable, built at full size for a demonstration artist. Yours is built from your own catalog, numbers, and plans.

See the services