Attribution

To generate attribution reports in Poplar, transactional (order) data must be shared with the platform by one of three ways:

Upload CSV Poplar's Shopify App Orders API

Attribution reports are updated as transactional data comes into our system. If you change your attribution window at any point after setting up your campaign, it will take time to refresh and reload your new attribution data.

To track and/or download a CSV record of your mailers after launch, see your campaign's History tab:

pageHistory

Understanding Your Reports

In-Home Dates

The attribution window begins when your mailings are in-home. To provide a direct comparison across postage types and holdouts, there is a slight nuance to this rule.

Let's say you sent two mailings on January 1:

Mailing A used Standard postage Mailing B used First Class postage

And let's say Mailing B was in-home on Jan 5 (e.g. Jan 5 was the Delivery Scan Date):

For all mail pieces from A and B, including any holdouts in A and B, we would consider January 5 as the in-home date. This allows you to directly compare the results from all mailings on January 1.

Reporting Window

You may select a reporting window to filter and see only mailings that were sent in that time period. In Poplar, reporting window refers to the date a mailing was sent (created manually or triggered). Please keep in mind that this is different from the attribution window.

To view reports by day instead of by week, simply adjust the Reporting Window to only show the past seven days:

Attribution Window

You can set a custom attribution window when you create a new campaign. By default, Poplar uses a 90-day attribution window. This is the period of time for which you wish to credit a customer's transaction to the mailing. We recommend a minimum of 30 days.

Before your attribution window has fully passed, you may begin to see some results, but because these are incomplete, we suggest waiting until the attribution window has fully passed before evaluating the performance of the mailings.

You can also update the attribution window of any campaign under Edit Campaign.

Changing the attribution window will trigger a refresh of a large quantity of data, which may take some time to load. If you shorten the attribution window, you'll also likely see fewer conversions for that campaign or send.

New Buyers

The Orders API allows you to specify whether transactions came from a new buyer (customer) or not. If you are using Poplar in any customer acquisition campaigns, or if you just want to filter your attribution reports by new customers, we recommend including this metadata through the Orders API.

We assume that all New Buyer data sent to our system is correct. Therefore, if Jane Doe made two orders and both orders were marked with the flagnew_buyer: true, your attribution reports would count Jane Doe as a new customer, twice.

Attribution Model

Poplar measures customer response to mailings using the last-touch attribution model. This means that we attribute orders to the most recent mail piece (within the attribution window).

Buyer Uniqueness

Per campaign, the first matching order from each customer is considered unique and any subsequent orders from the same customer that are also attributed to the campaign are not considered unique.

Note: A customer should only ever be reported as a New Buyer once, but per campaign you send out, each customer has a new opportunity to be considered unique.

We determine whether or not a customer is unique (per campaign) based on a couple of data points and the likelihood the data point is available.

The first data point we consider is the customer_id.
If customer_idis not provided, we default to using customer_email.
If neither data points are provided, we cannot determine the customer's uniqueness.
We don't use addresses to determine customer uniqueness as an address could map to multiple individuals.

Holdouts (A/B Testing)

Learn more about setting up A/B or multivariate tests.

Some attribution metrics in the reports are centered around the difference in performance between holdout vs. mailed groups. If you set your campaign to take holdouts, you'll see these additional metrics in your reports.

Data Refresh

Your attribution reports are updated as transactional data comes into our system. If you change your attribution window at any point after your set up your campaign, it will take some time to refresh and reload your new attribution reports.

Download Raw Matches

CSV files will contain all matches and attributions. If you only need attributions, please ensure you filter out matches. Under the column match_type

"A" means that the order was attributed to that particular mailing.
"M" means that the order simply matched to that particular mailing, but was not attributed to that mailing. We follow a last-touch attribution model across campaigns, so it could be that the mail piece was not the last one that the specific customer received.

Across All Campaigns

A CSV record of raw matches across all campaigns can be downloaded from the Transactional Data tab under API:

The download will not begin immediately. After a few minutes, you'll receive an email with the download link.

For a Single Campaign

A CSV record of raw matches for a specific campaign can also be downloaded from the campaign's Overview:

For campaigns using the Address Enrichment feature, your CSV download file will not include any address data generated through the matching process. We are only able to provide address data in the CSV download if the address came from your order data.

The download will not begin immediately. After a few minutes, you'll receive an email with the download link.

Common Questions

When should I upload my order data?

If you're planning to upload a CSV file, we recommend waiting until after your campaign has been in-home for 30 days before exporting data from your system and doing the upload.
If you're planning to send order data through the Orders API, we recommend setting up the API call as soon as you've sent your first campaign.
If you're planning to send data through the Shopify App, you should turn on the reporting when you set up your first trigger.

What happens if I upload an order that's a duplicate of an order I previously uploaded?

We define a duplicate order by any order that has the same exact order_id as an order previously uploaded into Poplar. The data attached to the most recent upload will replace the existing data.

For example, if you uploaded Order #12345 with an order total of $40.00 on March 1, but re-uploaded Order #12345 with an order total of $35.00 on March 5, Poplar would store the order total of $35.00.

Can I change the total for an order?

Yes! Here are two ways you can update your order:

Re-upload the same order with the new information. Make sure to include the same exact order_id as the order you're trying to update.
Use our Edit Order endpoint.

Can I delete my order data?

Yes, you can delete your order data under API > Transactional Data. Please note that this data deletion means you will lose all your attribution reporting.

If you want to delete an individual record, you can do so using the Delete Order endpoint.