# Meta Ads Attribution Settings 2026: 7-Day Click, 1-Day & Incremental | Enrich Labs

> Meta ads attribution settings in 2026: standard vs incremental, 7-day vs 1-day click, 1-day view, Jan 2026 API window cuts, and best settings by campaign type.

_Source: https://www.enrichlabs.ai/blog/meta-ads-attribution-settings-best-practices_

---

## TLDR

**Meta ads attribution settings** decide which conversions Ads Manager credits to your ads, and (for standard attribution) which click/view windows the delivery system optimizes toward. Meta currently offers **standard** and **incremental** attribution models at the ad set level ([Meta Business Help](https://www.facebook.com/business/help/460276478298895)). The common default pair is still **7-day click + 1-day view**, but that pair is not always the right reporting or optimization choice for leads, remarketing, or long B2B cycles.

**What changed in 2026:** Meta announced that on **January 12, 2026**, Ads Insights API `action_attribution_windows` stops returning **7-day view** and **28-day view**. Windows that remain available (aligned with Ads Manager) include **1-day click, 7-day click, 28-day click, 1-day engaged view, and 1-day view**, plus defaults from your ad set attribution settings ([Meta for Developers](https://developers.facebook.com/blog/post/2025/10/16/ads-insights-api-metric-availability-updates/)).

**Practical short answers:**

-   **Ecommerce product sales (short path):** start with **7-day click + 1-day view** (or 7-day click only if view-through inflates CPA vs your first-party truth).
-   **Lead gen / free trials / forms:** prefer **1-day click** (often **no view-through**) so organic and email conversions do not pile onto old ad impressions.
-   **Remarketing / warm audiences:** **1-day click** (tight window; they already know you).
-   **High-ticket B2B / long consideration:** **7-day click** (sometimes longer click windows where available); treat platform ROAS as directional, not board truth.
-   **Always:** pair Pixel + [Conversions API](https://www.facebook.com/business/help/AboutConversionsAPI), use `event_id` dedupe, and compare settings in Ads Manager before you change live optimization ([Meta Business Help: compare attribution settings](https://www.facebook.com/business/help/654970342692714)).

Helena at [Enrich Labs](https://www.enrichlabs.ai/) is an [AI marketing agent](/ai-marketing-agent) that can keep Meta execution, tagging, and conversion events consistent so your attribution windows describe real paths instead of broken tracking. For the full stack setup, see [Meta conversion tracking 2026 (Pixel, CAPI & first-party data)](/blog/setup-conversion-tracking-meta-ads-with-ai). For path models beyond Meta, see our [marketing attribution complete guide 2026](/blog/marketing-attribution-complete-guide-2026).

* * *

## What Meta ads attribution settings actually control

Attribution settings are not a vanity report filter only. In Meta Ads Manager you choose an **attribution model** and, for **standard** attribution, **attribution settings** (time windows and engagement types such as click, view, engaged view) at the **ad set** level. Those choices:

1.  **Inform ad delivery** (what the system optimizes toward under standard attribution).
2.  **Determine how conversions can be credited** to ads in reporting ([Meta Business Help](https://www.facebook.com/business/help/460276478298895)).

Two different conversations get mixed constantly:

Conversation

Question

Who cares

**Optimization attribution**

Which conversions should Meta bid toward?

Media buyers, automation agents

**Reporting attribution**

Which conversions should we _show_ when we judge creative or CPA?

Finance, founders, agencies

**Business truth**

Did this spend create incremental revenue or pipeline?

Boards, CMOs

Meta's UI answers the first two inside Ads Manager. The third needs first-party analytics, CRM, holdouts, and often [MMM / multi-touch outside Meta](/blog/marketing-attribution-complete-guide-2026).

**Definitions you need:**

-   **Click-through attribution:** a person clicked the ad, then converted inside the click window.
-   **View-through attribution:** a person saw the ad (impression), did not click (or the path is treated as view), then converted inside the view window.
-   **Engaged-view:** Meta's engaged-view concept (used in some video / engagement contexts) sits between a pure impression and a click; availability depends on objective and product surface. Treat it as a distinct engagement type when the UI offers it ([Jon Loomer's 2026 overview](https://www.jonloomer.com/meta-ads-attribution-2026/) summarizes how standard optimization ties to selected windows and engagement types).

If tracking is broken (no CAPI, weak Event Match Quality, missing `event_id` dedupe), **no attribution window saves you**. Fix measurement first: [Meta Pixel + CAPI setup](/blog/setup-conversion-tracking-meta-ads-with-ai).

* * *

## Attribution models in 2026: standard vs incremental

Meta documents two primary models you select when creating an ad set ([Meta Business Help](https://www.facebook.com/business/help/460276478298895)):

### Standard attribution

-   Optimizes delivery for the **time windows and user behaviors you select**.
-   Lets you choose whether conversions can be credited from **impressions, clicks, and/or video plays** (product-dependent).
-   This is the model most accounts still run day to day.
-   Example configuration language from practitioner docs: optimize for a selected window such as **7-day click, 1-day view** ([Jon Loomer](https://www.jonloomer.com/meta-ads-attribution-2026/)).

**Use standard when:** you want transparent windows, need to match historical reporting, or run volume campaigns where industry-default windows are good enough.

### Incremental attribution

-   Meta's product framing: credit conversions the ad is estimated to have **caused**, not every conversion that simply happened after an ad touch ([Meta Business Help](https://www.facebook.com/business/help/460276478298895)).
-   Goal: reduce over-crediting of conversions that would have happened anyway (especially brand-heavy and retargeting-heavy accounts).

**Use incremental when:** you care more about lift than raw conversion volume; you have enough conversion volume for Meta's model to learn; finance already distrusts view-through-heavy ROAS.

**Honest limits:** incremental is still a **platform estimate**. It is not a substitute for geo tests, PSA holdouts, or CRM pipeline contribution. For cross-channel board reporting, keep reading our [attribution control-tower guide](/blog/marketing-attribution-complete-guide-2026).

* * *

## Windows that still matter after the 2026 API cleanup

### What Meta deprecated for the Insights API

On **January 12, 2026**, Meta deprecated certain `action_attribution_windows` values on the Ads Insights API:

-   **7-day view (`7d_view`)** no longer returned
-   **28-day view (`28d_view`)** no longer returned

Windows that **continue** (consistent with Ads Manager behavior) include **1-day click, 7-day click, 28-day click, 1-day engaged view, and 1-day view**, plus defaults based on the advertiser's attribution settings, for fields like `actions`, `conversions`, and `results` ([Meta for Developers, Oct 16, 2025](https://developers.facebook.com/blog/post/2025/10/16/ads-insights-api-metric-availability-updates/)).

**Operator takeaway:** if your BI pipeline still requests `7d_view` or `28d_view`, rebuild it. Prefer **1-day view** as the only long-lived pure view window in many modern setups, and treat long view windows as historical debt.

### Windows you still compare in Ads Manager

Meta lets you **compare** reported conversions across settings including **1-day view, 1-day click, 7-day click, and 28-day click** (with product caveats) ([Meta Business Help: compare attribution settings](https://www.facebook.com/business/help/654970342692714)). Use compare mode **before** you change the optimization setting on a scaled ad set.

### Practical window menu (2026)

Window combo

What it emphasizes

Typical fit

**1-day click only**

Fast, high-intent paths; harsh on long consideration

Lead magnets, free trials, remarketing

**1-day click + 1-day view**

Short path + same-day view assist

App installs, local offers, impulse DTC

**7-day click only**

Week-long click paths; no view inflation

High-ticket, B2B demos, considered ecommerce

**7-day click + 1-day view**

Meta-style default for many sales campaigns

Broad ecommerce prospecting

**28-day click** (where available)

Very long click paths

Niche high-AOV; use sparingly; hard to defend vs other channels

**Incremental model**

Estimated causal credit

Mature accounts with volume + skeptical finance

Industry writeups still describe the **default configuration as 7-day click-through and 1-day view-through** for many sales setups ([Ryze AI overview of the same topic](https://www.get-ryze.ai/blog/meta-ads-attribution-settings-best-practices); treat competitor pages as SERP context, not gospel). Your job is to **match window to buying cycle**, not to copy a default forever.

* * *

## Best Meta ads attribution settings by campaign type (2026)

These are **starting points**, not laws. Always validate with Ads Manager compare columns + your first-party trial/purchase timestamps.

### 1\. Ecommerce / catalog sales (short AOV path)

**Start:** **7-day click + 1-day view** under standard attribution.

**Why:** Many shoppers click, leave, return within a few days. A pure 1-day click window under-credits legitimate multi-session carts. 1-day view can capture same-day assisted purchases without the old 7/28-day view hangover.

**Tighten when:** blended MER is healthy but Meta ROAS looks "too good" vs Shopify/Stripe. Drop view-through or move to **7-day click only**, then re-read CPA for 7-14 days.

**Stack:** Pixel Purchase + CAPI Purchase with shared `event_id`; product catalog IDs consistent ([CAPI overview](https://www.facebook.com/business/help/AboutConversionsAPI)).

### 2\. Lead generation (forms, magnets, webinar regs)

**Start:** **1-day click only** (no view-through).

**Why:** Free offers convert fast after a click. Long click windows + view-through often pull in people who would have signed up from email, SEO, or direct. Competitor guidance commonly flags inflated lead counts when 7-day windows stay on lead gen ([Ryze](https://www.get-ryze.ai/blog/meta-ads-attribution-settings-best-practices)).

**Optimize for:** the **qualified** event (SQL, booked call, trial start), not raw Lead, whenever volume allows. Enrich Labs default Meta conversion for trials is a **SUBSCRIBE / trial-start** style custom event in many accounts; match your CRM truth.

### 3\. Free trial / SaaS signup

**Start:** **1-day click** or **7-day click only** depending on trial friction.

-   Instant self-serve trial after ad click → **1-day click**.
-   Demo-assisted or multi-step onboarding before "trial started" fires → **7-day click only**.

**Avoid:** aggressive view-through on top-of-funnel thought-leadership ads if finance scores trials from CRM first-touch.

**Agent angle:** [Helena](/ai-digital-marketing-agent) can keep event naming, UTMs, and creative tests aligned so the window you pick describes one coherent funnel. See also [best AI for Meta ads 2026](/blog/best-ai-for-meta-ads-2026).

### 4\. App installs

**Start:** **1-day click + 1-day view** is a common mobile pattern (fast install-or-not behavior; app store discovery can be view-assisted the same day). Confirm against MMP (AppsFlyer, Adjust, etc.) and ATT reality: Adjust reported industry ATT opt-in (users shown the prompt) around **35%** in Q2 2025 data ([Adjust](https://www.adjust.com/blog/att-opt-in-rates-2025/)). Apple requires ATT before tracking via the advertising identifier in the described cases ([Apple](https://developer.apple.com/app-store/user-privacy-and-data-use/)).

### 5\. Remarketing / retargeting

**Start:** **1-day click only**.

**Why:** Warm users already visited. Long windows double-count email, SMS, and brand search. View-through on site retargeting is a classic ROAS mirage.

### 6\. High-ticket B2B / long consideration

**Start:** **7-day click only** (or longer click windows if your Meta surface still exposes them and your sales cycle justifies it).

**Why:** Path is sales-assisted. Platform conversion may be "Submit form," not closed-won. Report **sourced vs influenced** pipeline in CRM; do not let Ads Manager ROAS alone staff the board deck ([6sense B2B attribution context](https://6sense.com/science-of-b2b/2025-b2b-marketing-attribution-and-contribution-benchmark/) still shows marketers report ROI while journeys stay heavily anonymous).

### 7\. Video / engaged-view heavy campaigns

**Start:** standard attribution with **click + engaged-view** options **only if** the objective and UI expose engaged-view cleanly, and you have a plan to separate **brand** KPIs from **performance** KPIs.

**Rule:** never judge performance creative on engaged-view-inflated CPA without a click-only compare column.

### 8\. Brand awareness / reach

**Start:** do not force performance CPA windows onto pure reach. Use brand lift, site visitation studies, or MER movement. If you must pick a performance-adjacent window for a hybrid campaign, prefer **7-day click + 1-day view** and label reports as **directional**.

* * *

## How to change attribution settings without wrecking learning

1.  **Screenshot baseline:** CPA, cost per trial, frequency, and conversion count under current settings (7-14 days).
2.  **Use Compare attribution settings** in Ads Manager to see how counts move across 1d/7d click and 1d view **before** you edit the ad set ([Meta Help](https://www.facebook.com/business/help/654970342692714)).
3.  **Change one variable:** window **or** model (standard → incremental), not both on the same day on a scaled set.
4.  **Prefer a new ad set** when you need a clean learning read; heavy edits on live scaled sets muddle delivery history.
5.  **Hold budget steady** 5-7 days post-change (or through one full learning cycle) unless CPA is catastrophic.
6.  **Reconcile to first-party:** trials in Stripe/CRM vs Meta results within ± the window length.
7.  **Document the decision** in the ad set name (`…_7c1v_std` vs `…_1c_inc`) so agents and humans do not "fix" it later.

Automation note: if an [AI Meta ads agent](/blog/best-ai-for-meta-ads-2026) can edit bids but not windows, lock windows in a checklist. Helena's workflow bias is consistent events + creative iteration first; window thrashing is a last lever.

* * *

## Pixel, CAPI, EMQ: the floor under every window

Attribution settings distribute credit. **CAPI + Pixel** create the events worth distributing.

Minimum bar (2026):

1.  **Browser Pixel** for pageView + key events.
2.  **Conversions API** server events for the same purchase/lead/trial names ([Meta: About Conversions API](https://www.facebook.com/business/help/AboutConversionsAPI)).
3.  **Shared `event_id`** (and matching time) so Meta can dedupe browser + server.
4.  **Customer information parameters** (hashed email, phone, etc.) to lift **Event Match Quality** ([Meta: Event Match Quality](https://www.facebook.com/business/help/765081237991954)).
5.  **Test Events** tool QA before spend.
6.  **Primary conversion** aligned to business outcome (trial start > page view).

Deep setup: [Meta conversion tracking with AI (2026)](/blog/setup-conversion-tracking-meta-ads-with-ai). If the live page lags after a publish, hard-refresh or purge CDN cache on that slug.

* * *

## Meta numbers vs GA4 vs CRM (why they never match)

System

Typical lens

Common bias

**Meta Ads Manager**

Platform-attributed results inside your windows / model

Over-shares credit on view-through; walled garden

**GA4**

Data-driven or last-click style paths across sessions ([GA Help](https://support.google.com/analytics/answer/10596866?hl=en))

Under-sees walled garden; consent mode gaps

**Google Ads**

Defaults many conversion actions to **data-driven** after retiring several rule-based models ([Google Ads Help](https://support.google.com/google-ads/answer/6259715?hl=en))

Google-heavy last paths

**CRM**

Opportunity sourced/influenced, sales cycle reality

Slow; incomplete UTM hygiene

**Rule for operators:** pick **one system of record per decision**.

-   **In-channel bid changes** → Meta (with a window you trust).
-   **Cross-channel budget** → first-party + MER / pipeline.
-   **Causal claims** → experiments / MMM.

Braze's framing still holds: path models weaken when journeys fragment and privacy cuts identity ([Braze on attribution challenges](https://www.braze.com/resources/articles/challenges-of-marketing-attribution)).

* * *

## Common Meta attribution mistakes (and fixes)

1.  **Using 7-day click + 1-day view on pure lead magnets** → inflated leads. Fix: 1-day click only.
2.  **Judging remarketing on long windows** → double-count. Fix: 1-day click; separate prospecting reports.
3.  **Changing windows weekly** → noise. Fix: 2-week minimum reads.
4.  **No CAPI** → iOS and browser loss. Fix: server events + EMQ.
5.  **Optimizing for the wrong event** → cheap leads, empty pipeline. Fix: trial/purchase quality events.
6.  **BI still pulling `7d_view` / `28d_view` after Jan 12, 2026** → broken pipelines ([Meta Developers](https://developers.facebook.com/blog/post/2025/10/16/ads-insights-api-metric-availability-updates/)). Fix: rebuild windows list.
7.  **Treating incremental attribution as MMM** → category error. Fix: label it platform-estimated lift.
8.  **Ignoring engaged-view in video** → surprise CPA. Fix: compare click-only columns.
9.  **UTM chaos** → CRM cannot reconcile. Fix: locked taxonomy; agents enforce it.
10.  **One global setting for every ad set** → lazy. Fix: matrix by objective (table above).

* * *

## How to test attribution settings the clean way

### A. Offline compare (no delivery change)

Use Meta's **compare attribution settings** on historical delivery ([Help](https://www.facebook.com/business/help/654970342692714)). Chart:

-   Results @ 1d click
-   Results @ 7d click
-   Results @ 1d view
-   Your live optimization setting

If 1d view is 40%+ of "results," your ROAS story is impression-heavy. Decide consciously.

### B. Split ad sets (delivery change)

Duplicate ad set → new attribution setting → equal budget → 7+ days → compare **cost per first-party conversion**, not only Meta CPA.

### C. Incrementality layer

For six-figure monthly Meta spend, add geo split or conversion holdout. Platform incremental attribution is a hint; it is not the experiment.

### D. Agent-assisted monitoring

Have your operator (human or [Helena](https://agent.enrichlabs.ai/marketing/register)) alert when:

-   Meta CPA diverges >30% from first-party CPA for 3 days
-   EMQ drops
-   View-through share spikes after a creative shift

* * *

## Decision matrix: pick a setting in 60 seconds

If your primary KPI is…

And path length is…

Start with…

Purchases, AOV mid, DTC

1-7 days

7-day click + 1-day view

Purchases, high AOV

7-30 days

7-day click only (test 28-day click carefully)

Leads / magnet

Same day

1-day click only

Trial start (self-serve)

Same day–3 days

1-day click or 7-day click only

Pipeline (B2B)

Weeks

7-day click only + CRM influence

Remarketing ROAS

Hours–1 day

1-day click only

App install

Same day

1-day click + 1-day view

Finance wants lift

Any

Test incremental model + holdouts

Then **lock** the choice in naming + docs so the next person does not "optimize" it on a whim.

* * *

## FAQ

### What is the best Meta ads attribution setting in 2026?

There is no single best setting. **7-day click + 1-day view** remains a common sales default; **1-day click only** is usually better for leads and remarketing. Choose by conversion lag and whether view-through is defensible against your first-party data ([Meta model overview](https://www.facebook.com/business/help/460276478298895)).

### Is 7-day click or 1-day click better?

**1-day click** is stricter and usually better for fast offers (leads, many trials). **7-day click** fits multi-session ecommerce and longer B2B form fills. Compare both in Ads Manager before you switch optimization ([compare settings](https://www.facebook.com/business/help/654970342692714)).

### Should I use view-through attribution?

Use **1-day view** sparingly and knowingly. Long view windows (7d/28d view) are gone from Insights API responses as of the Jan 12, 2026 change ([Meta Developers](https://developers.facebook.com/blog/post/2025/10/16/ads-insights-api-metric-availability-updates/)). View-through helps some app and impulse DTC cases; it pollutes lead gen and remarketing.

### What is incremental attribution on Meta?

A model that aims to credit conversions the ad is estimated to have **caused**, versus standard window-based credit ([Meta Help](https://www.facebook.com/business/help/460276478298895)). Still not a full MMM replacement.

### Why did my API reports break in 2026?

If you requested `7d_view` or `28d_view` on Ads Insights, those windows were deprecated Jan 12, 2026 ([changelog context via Meta Developers post](https://developers.facebook.com/blog/post/2025/10/16/ads-insights-api-metric-availability-updates/)). Update integrations.

### Do attribution settings change delivery or only reports?

For **standard** attribution, settings inform **delivery and credit**. Reporting compare tools can show alternate windows without always implying you changed optimization. Read Meta's model copy carefully when you edit an ad set ([Help](https://www.facebook.com/business/help/460276478298895)).

### How do iOS and privacy affect Meta attribution?

Browser and IDFA paths are incomplete. ATT opt-in rates remain limited in industry datasets ([Adjust 2025](https://www.adjust.com/blog/att-opt-in-rates-2025/)). CAPI + first-party identifiers matter more than window micro-tweaks.

### Should every ad set use the same attribution setting?

No. Match **objective and funnel stage**. Prospecting ≠ remarketing ≠ lead magnet.

### How does this relate to GA4 data-driven attribution?

Different garden, different model. GA4 documents data-driven and last-click style options for key events ([GA Help](https://support.google.com/analytics/answer/10596866?hl=en)). Expect mismatch; reconcile with a primary KPI definition.

### Can an AI agent manage attribution settings?

Yes as **checklist + monitoring** (drift alerts, EMQ, first-party vs Meta CPA). Window changes still need a human policy. [Enrich Labs / Helena](https://www.enrichlabs.ai/) focuses on execution consistency so settings stay interpretable. Try the [3-day trial](https://agent.enrichlabs.ai/marketing/register).

### Where do I learn full-funnel attribution beyond Meta?

[Marketing attribution complete guide 2026](/blog/marketing-attribution-complete-guide-2026) covers MTA, MMM, CRM sourced/influenced, and 90-day rollout.

### What about conversion tracking setup itself?

[Meta conversion tracking 2026: Pixel, CAPI & first-party data](/blog/setup-conversion-tracking-meta-ads-with-ai).

* * *

## Bottom line

**Meta ads attribution settings best practices in 2026** are simple to say and hard to keep:

1.  Fix **Pixel + CAPI + EMQ** before you debate windows.
2.  Default sales paths to **7-day click ± 1-day view**; default leads/remarketing to **1-day click**.
3.  Use **compare mode** before you touch a scaled ad set.
4.  Rebuild any BI that still calls **7d/28d view** after Meta's Jan 2026 Insights API cleanup.
5.  Treat **incremental** as estimated lift, not board-grade causality.
6.  Keep one **system of record** per decision (Meta vs GA4 vs CRM).

When execution, events, and taxonomy stay consistent, windows become a strategic choice instead of a weekly argument. That is the job [Helena](/ai-marketing-agent) is built for inside [Enrich Labs](https://www.enrichlabs.ai/), from **$39/mo** with a [3-day free trial](https://agent.enrichlabs.ai/marketing/register).

**Related:** [Best AI for Meta ads 2026](/blog/best-ai-for-meta-ads-2026) · [Meta campaign objectives explained](/blog/meta-ads-campaign-objectives-explained) · [Best MCP for Meta ads](/blog/best-mcp-for-meta-ads) · [Marketing attribution guide](/blog/marketing-attribution-complete-guide-2026)