SKAN Conversion Values Setup
Set up SKAN conversion values using a decision framework. Map events, windows, MMP data, and revenue before Apple Search Ads optimization.
Recommended
Find Profitable Apple Search Ads Keywords
Feeling lost with Apple Search Ads? Find out which keywords are profitable 🚀
The short answer: Pick a SKAN conversion value mapping shape based on your specific app business model rather than importing generic templates.
SKAN conversion values setup is where Apple Search Ads measurement either becomes useful or turns into a spreadsheet séance. The goal is not to encode every tap, screen, subscription state, and product event into one magical number. The goal is to choose a small set of early signals that help you compare privacy-safe cohorts without pretending SKAN is a user-level revenue database.
Use this worksheet when Apple Search Ads is already running, SKAN or MMP reporting exists, and the team needs a safer way to decide what a conversion value should mean. It is built from repo-internal Apple Search Ads guides on SKAN optimization, attribution limits, tracking, and rules. No outside CPT, CPI, CPA, ROAS, LTV, install-rate, revenue, or retention targets are assumed. If a threshold matters, it needs to come from the account’s own product and revenue data.
Quick answer
Set up SKAN conversion values by mapping a few account-specific product milestones to clear value buckets, documenting the measurement window, reconciling the result with MMP and internal revenue data, and freezing bid automation when the map is stale or incomplete. Apple Search Ads should own spend, taps, and Apple-reported installs. SKAN should own delayed cohort direction. The MMP should reconcile downstream events. Internal revenue should decide whether the cohort is actually worth scaling.
A useful setup is boring on purpose. It answers: what event does this value represent, when can the postback be read, which campaign theme does it describe, and what action is allowed after the window completes?
Conversion-value setup worksheet
Create one row for each business milestone that might influence Apple Search Ads decisions. Keep the row simple enough that a marketer, data owner, and product owner can all understand it.
| Field | What to define | Why it matters |
|---|---|---|
| Business model | Subscription, one-time purchase, lead form, marketplace action, or ad-supported app | The useful early signal depends on how the app earns |
| Product milestone | Activation, account creation, onboarding completion, trial start, purchase intent, purchase, renewal proxy, or key retained action | Conversion values should represent business progress, not vanity clicks |
| SKAN value or bucket | The value, range, or coarse bucket used for that milestone | The team needs one shared meaning for each signal |
| Event owner | SDK, server event, MMP mapping, or internal warehouse field | Ownership prevents mystery numbers from driving spend |
| Measurement window | The postback, event, or revenue window used before review | Delayed SKAN signals should not be judged like live keyword rows |
| Apple Search Ads lane | Brand, exact, broad, Search Match, competitor, category, or discovery | Mixed lanes make cohort reads muddy |
| MMP reconciliation | Matching event, attribution window, or cohort label | Helps catch mapping drift and channel-window mismatches |
| Revenue check | The internal product or billing field that confirms quality | Install quality is not enough when revenue arrives later |
| Allowed action | Freeze, hold, inspect, lower, cap, split, promote, or scale gradually | A map is only useful if it controls the next move |
| Review owner | Person or team responsible for approving changes | Makes rollback possible when the next window disagrees |
The worksheet should live next to the rule or dashboard that uses it. If the bid rule cannot name the conversion-value map and the window it depends on, the rule is not ready to run.
Pick the right mapping shape
Different apps need different SKAN conversion-value shapes. Do not copy a map from another app because it looked tidy in a conference slide. That is how attribution theater gets a costume budget.
| App situation | Better mapping shape | What to avoid |
|---|---|---|
| Subscription app with trial | Activation, trial start, paid start, and early renewal proxy | Treating install volume as the whole value signal |
| One-time purchase app | Purchase intent, checkout start, purchase, and refunded or cancelled order flag if available internally | Overweighting lightweight browsing events |
| Lead-generation app | Qualified form start, completed lead, verified lead, and downstream sales handoff | Counting every signup as equal when lead quality differs |
| Marketplace app | Buyer activation, seller activation, first transaction, and repeat engagement proxy | Mixing buyer and seller milestones in one unreadable bucket |
| Ad-supported or engagement app | Onboarding completion, first meaningful session, retained session proxy, and ad inventory eligibility | Encoding too many low-value engagement events |
| Low-volume campaign lane | Coarser buckets by theme or campaign group | Chasing single-keyword conversion values below useful signal levels |
The SKAN optimization source notes are clear about source ownership: Apple Search Ads reports can tell you whether a lane delivered impressions, taps, spend, and installs. SKAN can tell you whether an aggregated cohort is moving in a useful direction. The MMP and internal revenue data decide whether the map still lines up with downstream quality.
Data readiness checklist
Before letting conversion values influence bids or budgets, check the setup like this:
- Confirm the product funnel. List the current onboarding, activation, purchase, subscription, renewal, or lead-quality milestones.
- Choose only useful early events. Pick signals that happen early enough to inform campaigns but late enough to indicate real value.
- Document the SKAN window. Write down when the postback or conversion-value read is complete enough for review.
- Separate traffic lanes. Do not mix brand, exact, broad, Search Match, competitor, and discovery traffic in the same read if they need different decisions.
- Match the MMP window. Note where the MMP uses a different attribution or event window from Apple or SKAN.
- Add an internal revenue check. Use product, billing, refund, renewal, or server-side data as the final quality read when available.
- Define freeze conditions. Freeze automation when Apple, SKAN, MMP, API, or revenue pulls are stale, partial, or mapped to different windows.
- Log every mapping change. Record the owner, date, reason, expected effect, and rollback condition.
This is not busywork. A stale conversion-value map can make a weak cohort look promising, or make a good cohort look dead because the product changed and the map did not.
Setup sequence for Apple Search Ads teams
Use this order when building or repairing the map:
| Step | Action | Output |
|---|---|---|
| 1 | Inventory campaign lanes and keyword themes | Brand, exact, broad, Search Match, competitor, category, and discovery are separated |
| 2 | List product milestones by business model | A small set of candidate events with owners |
| 3 | Choose the first mapping version | A documented value or bucket map with no mystery events |
| 4 | Connect SKAN, MMP, and internal windows | A row showing when each source can be read |
| 5 | Back-check recent cohorts | Directional evidence that the map describes real product behavior |
| 6 | Attach allowed actions | Freeze, hold, inspect, lower, cap, split, promote, or gradual scale |
| 7 | Add alerts | Stale-data and mismatched-window checks before automation acts |
| 8 | Review after product changes | Mapping update whenever onboarding, pricing, purchase flow, or event tracking changes |
Keep the first version narrower than your ambition. A conversion-value map that everyone can explain beats a dense one that requires an archaeologist with admin access.
What should not move bids automatically
Do not let these signals trigger automatic Apple Search Ads changes by themselves:
| Signal | Why it is risky | Safer move |
|---|---|---|
| Fresh Apple installs but incomplete SKAN window | Install volume arrives before cohort value | Hold until the chosen window completes |
| Strong SKAN value but stale MMP data | Downstream event mapping may be incomplete | Freeze and reconcile before scale |
| MMP event lift but no internal revenue read | Event quality may not match payback | Hold or inspect until business data arrives |
| Conversion-value map changed mid-window | Cohorts are not comparable | Split the review before and after the mapping change |
| Broad and brand traffic in one read | Brand demand can flatter broader traffic | Separate lanes before judging quality |
| Low-volume keyword row | Privacy thresholds and aggregation can hide signal | Review by theme, campaign, or country group |
These guardrails intentionally avoid universal thresholds. Your acceptable CPA, payback window, and value bucket definitions belong to the account. Copying someone else’s target is just outsourcing your future postmortem.
Decision Matrix
| Scenario | Recommendation | Why |
|---|---|---|
| Subscription app with free trial | Map values to activation, trial start, paid conversion, and early renewal proxy. | Prevents treating install volume as the entire value signal when payback arrives later. |
| One-time purchase app without recurring revenue | Map values to purchase intent, checkout start, purchase, and a cancelled or refunded order flag. | Avoids overweighting lightweight browsing events that do not correlate with actual transactions. |
| Lead-generation app with variable lead quality | Map values to qualified form starts, completed leads, verified leads, and downstream sales handoffs. | Stops counting every signup as equal when downstream lead quality and sales conversion differ significantly. |
| Marketplace app with buyer and seller sides | Map values to buyer activation, seller activation, first transaction, and repeat engagement proxy. | Prevents mixing buyer and seller milestones into one unreadable aggregated bucket. |
| Low-volume campaign lane or new country group | Use coarser conversion value buckets organized by theme or campaign group. | Chasing single-keyword conversion values below useful signal levels introduces noise from privacy thresholds and aggregation. |
Recommended Next Step
If the map is missing owners, windows, or revenue checks, start with the Apple Search Ads Attribution Limits reconciliation checklist before changing bids. If the map is documented and the windows are complete, connect it to the Apple Search Ads SKAN Optimization worksheet and the Apple Search Ads Rules and Alerts workflow so stale data freezes automation instead of quietly spending money.
Further Reading
Start Here
Decision Pages
Tools and Calculators
FAQ
How many SKAN conversion values should I use?
Use as many as the account can explain, reconcile, and review consistently. A small set of milestone buckets is often safer than a dense event map.
Should SKAN conversion values map directly to Apple Search Ads keywords?
Usually no, because SKAN is delayed and aggregated, making it safer to read values by campaign lane or cohort window. Use Apple Search Ads reports for keyword delivery and SKAN, MMP, and revenue data to judge downstream quality.
When should I update the conversion-value map?
Review the map after major onboarding, pricing, trial, purchase, event-tracking, or revenue-model changes. Also review it when Apple, MMP, and internal revenue disagree in a way that timing or attribution windows cannot explain.
Can I use outside CPI, CPA, or ROAS targets in the setup?
Not for real account decisions, because bid logic should use the app’s own spend, installs, event quality, revenue, refund, renewal, and payback data. Outside numbers can teach the concept but copying them turns the map into a decorative dashboard accessory.
Frequently Asked Questions
Why should you avoid copying generic SKAN conversion value templates?
What SKAN conversion values should be mapped for a subscription app?
What elements need to be documented when setting up SKAN conversion values?
How do SKAN postbacks work alongside MMP and internal revenue data?
Sources & Citations
Next step
Find Profitable Apple Search Ads Keywords
Feeling lost with Apple Search Ads? Find out which keywords are profitable 🚀