Attribution model

Key features of the attribution model

• Email revenue can be attributed on opens as well as clicks. Opens are off for existing accounts until you turn them on.
• Each interaction type — email opens, email clicks, SMS clicks, SMS deliveries — has its own attribution window. All other channels share one window.
• You can set windows in days or hours, up to 30 days (720 hours).
• You can exclude one or more campaigns from attribution.
• Channels are split into three groups — campaigns, website, and in-app — and each group is attributed separately. Within a group, an order is attributed only once.
• Deleting customer data doesn’t affect revenue metrics.
• Lead generation forms are excluded from attribution — orders are never attributed to them.

What is attribution

• Attribution is a set of rules that decides whether a target action happened because of your marketing campaigns — and if so, which ones.
• A goal is the customer action you want more of. You launch campaigns to move this metric.

Use attribution to see how well each channel and campaign contributes to your goals, so you can decide how to split your marketing budget across them.

How the attribution model works

Maestra Platform uses the last-non-direct-touch model.

Channels are split into three groups, and orders are attributed in each group separately:

• Campaigns — email, SMS, web push, and other messaging channels
• Website — forms and pop-ups
• In-app

Within a group, only one interaction gets the order — the last non-direct one, following the rules below.

Say a customer clicks a web push, then clicks an email, then places an order. Both are campaigns, so only the email gets the order.

Across groups, the same order can be attributed more than once.

Say a customer fills out a website pop-up, then clicks an email, then places an order. The pop-up and the email are in different groups, so each of them gets the order. That way, reports show how much each group earned.

Because one order can count in several channel groups, the attributed revenue of all groups added together can be higher than your actual revenue.

Which interactions count

Not every interaction can anchor attribution — it depends on the channel:

Channel	Interactions that can anchor attribution
Email	Opens and/or clicks. You can turn each one on or off in the settings; at least one must stay on.
SMS and other SMS-like channels (e.g., MMS)	Clicks and/or deliveries. You can turn each one on or off; at least one must stay on.
Web push, mobile push, Viber, website pop-ups and forms, in-app	Clicks only

When an order is attributed to a campaign

An order is attributed to a campaign when all of these are true:

• The order matches the goal’s settings — status and first-action touchpoint.
• The interaction happened within the attribution window for that interaction type. The window is counted back from the order to the interaction (see below).
• There were no other website visits between the campaign interaction and the order’s first action. Direct visits (typing the URL into the address bar or using a bookmark) and visits with utm_source=maestra don’t break the chain.

Orders are never attributed to:

• Campaigns excluded in the attribution settings
• Lead generation forms and pop-ups

Deleted data doesn’t affect revenue metrics or conversion.

How the attribution window works

The window is always counted back from the order to the interaction itself — not from the moment the message was sent. The question the model asks is: “Did the customer open or click this message (or, for SMS, receive it) within the last X days before the order?”

Each interaction type has its own window. Say you set 5 days for opens and 10 days for clicks. An order is attributed to an email if the customer opened it within 5 days before the order or clicked it within 10 days before the order.

There’s no precedence between an open and a click. An open counts exactly like a click:

• Same last-non-direct-touch model
• Same rule about visits from other sources in between
• Same list of excluded campaigns
• Same goals

Every interaction is checked against its own window. Among all the interactions that qualify, the most recent one gets the order — a click 6 days before the order doesn’t beat an open 1 day before it.

Because the window starts at the interaction, not the send, there’s no limit on how old the message itself can be. If a customer opens a two-month-old newsletter and orders the next day, that newsletter gets the order — as long as nothing more recent qualifies.

Comparing channels

Different interaction types can have different windows — for example, 5 days for email opens, 10 days for email clicks, and 5 days for SMS deliveries. A channel with a wider window collects more orders for structural reasons, not because it performs better.

Keep this in mind when you read channel rows side by side in the Revenue & Drivers report: attributed revenue isn’t comparable across channels like-for-like unless you’ve set identical windows.

Attribution settings

Go to Settings → System → Goals and Attribution and open the Attribution tab.

The page has four blocks: Settings, Email, SMS, and Other channels.

Excluded campaigns

In the Settings block, use Excluded campaigns to list transactional and other campaigns that don’t affect conversion and shouldn’t take revenue away from your marketing campaigns — for example, an SMS with an authorization code at checkout or a double opt-in email.

Excluded campaigns never anchor attribution, in any channel.

You can exclude specific campaigns, or a set of campaigns by filter (tag, channel, type, and so on). New campaigns that match a filter are excluded automatically. You can exclude up to 100 campaigns and up to 100 filters.

Email

• Clicks — counts an order if the customer clicked a link in the email within the window before the order.
• Opens — counts an order if the customer opened the email within the window before the order.

Each row has its own toggle and window. At least one of the two must stay on.

Turning on opens raises attributed revenue — usually noticeably, because opens catch customers who read the email and then go to your website directly without clicking. That’s expected, not a bug.

Excluding opens

Under the Opens row, two toggles decide which opens get thrown out of attribution:

• Exclude Apple MPP opens from attribution — removes every open that came through Apple’s Mail Privacy Protection (MPP) proxy.

  Apple pre-loads email images for its Mail users, so some MPP opens are machine-generated. But real customers who read the email in Apple Mail come through the same proxy with the same signature, and there’s no way to tell them apart. So turning on this toggle drops real readers too: their orders only get attributed if they also clicked. Expect attributed revenue to go down when you enable it.

• Exclude bot opens from attribution — removes opens generated by security scanners and corporate mail filters.

Both toggles are off by default, so bot opens and Apple MPP opens are counted. Both affect attributed revenue only — the open rate in reports doesn’t change.

SMS

• Clicks — counts an order if the customer clicked a link in the SMS within the window before the order. Requires link shortening and click tracking.
• Deliveries — counts an order if the SMS was delivered within the window before the order. Use this if you send SMS without links or don’t track clicks — customers can act on the text of an SMS without clicking anything.

You can turn on both at the same time; at least one must stay on. SMS has no open event, so delivery is the closest engagement signal — windows here are usually shorter than for email.

Other channels

Web push, mobile push, Viber, website pop-ups and forms, and in-app are attributed on clicks, with one shared Attribution window.

Once you click Save, changes apply from the next attribution calculation (see the update schedule below). Every change is recorded in the account event log.

Goal settings

Goal settings live on the next tab:

Every account has a default goal, Orders, created automatically. It includes orders in the Placed, Paid, and Delivered status categories across all touchpoints. You can create up to 10 goals.

Update schedule:

• Last 30 days of data — daily, by 09:00.
• Last 2 years of data — every Monday by 09:00, if you’ve created a new goal or edited the attribution settings.

Where to find attributed revenue

Attributed revenue shows up as one number — it isn’t split into open-attributed and click-attributed revenue.

Reports

Use these reports to see how channels and individual campaigns contribute to your goals.

• Revenue & Drivers report — track your goals and how each channel affects them. Add a target to see attributed revenue for every channel.

  

• Campaigns report — see how each campaign affected the goal you’ve chosen.

  

• Website personalization report — see how each personalization campaign affected the goal you’ve chosen.

  

Filters

Use filters to select specific orders attributed to campaigns.

• Attributed to campaign filter:

  

• Attributed to flow filter:

  

Export

Add the attribution flag to your export settings, and the file will include the campaigns that drove each conversion.

Why data from an external analytics system might not match Maestra Platform reports

1. External analytics tools may not account for canceled and returned orders — their stats use the order value right after the order is placed. In Maestra Platform, you can exclude these orders by setting up goals. Excluded orders don’t appear in reports and don’t count toward final revenue.
2. If you have a loyalty program connected, compare the amounts sent to Maestra Platform and to your external analytics tool. Maestra Platform receives the order value after promotions; other systems may be set up differently.
3. If the gap is large, check your integration — the tracker may not be installed correctly, or there may be fake orders (an order far above your average order value). Make sure the Maestra tracker is installed on your website.
4. If Maestra Platform shows fewer orders than your site database or GA, check your account’s integration: order data you send to GA should also reach Maestra Platform, and vice versa. Make sure no orders are placed without Maestra Platform knowing about them — that is, the order create and update operations are called for every order. For example, an order placed through a pop-up sends data to GA, but the operation that passes it to Maestra Platform isn’t set up.
5. Attribution windows and interaction types differ between systems. Maestra Platform lets you choose which interactions count (opens, clicks, deliveries) and set a separate window for each interaction type, up to 30 days. Another system may count from the send date, use different windows, or filter Apple MPP and bot opens differently. Compare the settings before you compare the numbers.

FAQ

I turned on opens and attributed revenue jumped. Is that a bug?

No. Opens catch the large group of customers who read an email and then go to your website directly without clicking a link. Until now, those orders weren’t attributed to email at all.

How much revenue shifts depends on your audience. Accounts with lots of Apple Mail readers see a bigger uplift, because Apple MPP opens are counted by default.

Why did an old campaign suddenly get orders attributed to it?

The window is counted from the interaction, not from the send date. If a customer opened an old email (or Apple Mail pre-loaded it) shortly before ordering, and nothing more recent qualified, that old email gets the order. This affects the per-campaign breakdown more than the total.

Should I exclude Apple MPP opens?

Only if you’re fine with real Apple Mail readers no longer counting either — there’s no way to tell them apart from Apple’s pre-load. Most platforms that offer this setting also keep it off by default. If you need parity with another tool, match its setting.

Is my 5-day window the same as a 5-day window in another platform?

Not necessarily. Some platforms count the window from the moment the message is received; Maestra Platform counts it from the interaction. A window counted from the interaction is slightly looser, so the same number of days gives slightly higher attributed revenue.

Does changing the settings recalculate the past?

Settings apply from the next attribution calculation. Attributed revenue for earlier periods is recalculated during the regular 2-year update — but email opens are only collected for attribution from the moment you turn them on. Orders from before that date can’t be re-attributed to opens, so expect a step in your charts from the day you switch opens on.

Will SMS clicks still be collected if I attribute SMS on deliveries only?

Yes. Clicks are still tracked and shown in reports — they just don’t anchor attribution.

Does the open rate change when I exclude bot or MPP opens?

No. These toggles affect attributed revenue only. Open and click metrics in reports stay the same.