Skip to main content

A Guide to Black Crow Attribution

Front page

The purpose of this page is to give you a quick look at your shop’s KPIs, with an eye toward yesterday’s performance, and how that compares to the past week. You can select any date range you like, of course, but typically you’ll be checking this page to see if there are any recent trends you might want to investigate: a spike in revenue, a decrease in AOV, an unusual number of refunds. If there are any KPIs that are important to your business that you don’t see here, please email those to david.peer@blackcrow.ai; we’d love for this page to be as comprehensive as possible.

Acquisition

This page is designed for paid marketers, and empowers them with a variety of performance measures. No one attribution methodology is perfect, so Black Crow offers a few perspectives from different data sources so paid marketers can reach their own conclusions about what’s working.

Web analytics

This is a high-level view of the volume and quality of the traffic that you’re driving to your website. Note that Black Crow has superior user identification technology to GA, so our numbers (particularly our new user rate) may disagree somewhat with theirs. But we think this makes our numbers more accurate.

Marketing platforms

Here we see specifically how your paid marketing platforms are performing in driving would-be customers to your website. If there is a paid marketing platform you advertise on that you don’t see included here, please connect it in our portal's Integrations page.

Attribution

Black Crow's attribution methodologies combine data gathered from user behavior on your website and data from your connected marketing platforms. The most important type of data we gather from your website is clicks: visits that were referred by a marketing platform, with accompanying utm parameters that offer important context such as the referring campaign, ad ID, etc.

For this reason it's critical that clicks on your ads can be tracked, which requires that ads be comprehensively tagged with utm parameters. Be sure to add our utm parameters to every ad you have running in every ad platform. The parameters below begin with an ampersand (“&”) so that you can easily add them to whatever parameters you already have, but if an ad doesn’t already have utm parameters populated, be sure to remove that first ampersand:

  • Meta: &bc_source={{site_source_name}}&bc_campaign={{campaign.id}}&bc_adid={{ad.id}}
  • Google Ads: &bc_source=google&bc_campaign={campaignid}&bc_adid={creative}
  • TikTok: &bc_source=tiktok&bc_campaign=__CAMPAIGN_ID__&bc_adid=__CID__
  • Pinterest: &bc_source=pinterest&bc_campaign={campaignid}&bc_adid={adid}

We’ll read off of any data that we find in your utm parameters, even if we’re not connected to that platform, so if your affiliate platform has links tagged with utm parameters, you’ll see those clicks accounted for in the “Channels” table.

Attribution methodologies

Black Crow distributes credit for an order across any marketing channels it finds in the customer journey prior to purchase. How that credit is distributed depends on the attribution methodology, but in every case an order is accounted for only once; for example, Meta might get credit for 60% of a given order, and Google Ads the remaining 40%. This approach differs from Triple Whale, which allows full credit for an order to be attributed to multiple platforms. Black Crow supports four attribution methodologies:

  • First touch: This methodology is purely clicks-based, and gives full credit to the first channel the customer clicked on an ad for, regardless of any clicks that came afterward.
  • Last touch: This methodology is purely clicks-based, and gives full credit to the last channel the customer clicked on an ad for, regardless of any clicks that came before.
  • Even weight: This methodology is purely clicks-based, and gives even credit to every channel the customer clicks on, regardless of where those clicks fall in the customer journey.
  • Discovery weighted: This methodology is grounded in clicks, but integrates data from ad platforms to more accurately credit discovery channels such as paid social.

Customers

The goal of this page is to help you understand your customers: what they’re buying, what discounts they’re taking; what appeals to new customers and what appeals to your best customers.

Order analytics

The critical thing to understand about this view is that an order can be counted more than once: if a customer buys products A and B in the same order, for instance, or uses two different discount codes, that customer’s order will be counted in both of the appropriate rows of the table. Note the inclusion of Black Crow’s LTV predictions here as well: this will indicate which of your products are favored by your highest-value customers, or perhaps what effects discount codes are having on LTV.

LTV

The purpose of this view is to understand the evolution of your customers’ LTV by different cohorts. A
“New customer cohort” is a group of new customers: those who became customers in a certain month, say, or who all bought a certain product in their first order. You can choose which cohorts to examine in the “Group by” selector. A “Timeframe” is a period of time over which you can discretely measure the progress of a cohort’s LTV. We support week, month, quarter, and year timeframes. Let’s look at an example with fake dataLet’s start with the “2024-01” row, the fifth one up. Because we’ve selected “Month” as our group by, the cohort here is all new customers in the month of January 2024. The timeframe is also “Month”, so the x-axis is a count of months. The “0” timeframe is always AOV of the customers’ first orders. So in January of 2024, our new customers on average spent $277.43 in their first order. After 1 month their LTV had increased to $301.75; that is, they spent an additional $24.32 in subsequent orders, on average, after their first order. After two months their LTV had increased to $308.19. This continues on up through the 5-month timeframe, after which the data ends because at the time of writing (June 2024), we only have 5 months of history on new customers from January of 2024. Data for subsequent timeframes will be revealed as time goes on.

A few insights from this chart of the sort you may wish to look for in your store’s data:

  • AOV was low in November of 2023 (Black Friday/Cyber Monday shoppers?), and the LTV of this cohort was correspondingly low as well.
  • AOV is very low in May of 2024, and it looks like we should expect far less LTV from this cohort than prior cohorts as well.
  • It seems that LTV was markedly higher prior to November of 2023. This may be that we had a smaller, more loyal customer base then, one which we’ve since expanded on, to our benefit. Total sales may have more than compensated for the drop in LTV. But we may wish to tease out what happened here.

One last note: LTV does not always increase with each timeframe, especially when grouping by cohorts that aren’t time-based, like product, or marketing channel. This is because these cohorts may be composed of customers from different time periods, when your customers’ LTV profile was quite different. For example, if you’re looking at customers referred by Facebook, and your LTV has increased dramatically since a year ago, you may find that the LTV of your Facebook cohort appears to decrease over time, since the customers who are eligible for longer timeframes are older customers, and customers eligible for shorter timeframes are those same older customers but many newer customers as well. If you see this happening, you may wish to control for the effects of time with the date range filter, say by focusing on single months or quarters at a time.

Was this article helpful?