22 Things Bamboozling Your Marketing Attribution Data
I've painstakingly orchestrated attribution/acquisition initiatives on several occasions in the past, with varying degrees of success, and having to explain their complexity to executives is probably my least favorite marketing task.
Here's a blog post you can reference as a marketer when talking to your various executives that outlines why these initiatives are so hard, if not impossible, to execute.
Phase 1: Data Collection Issues
Let's just take it from the top of the funnel I guess, and start this painful journey.
1) The Shadow Funnel
Sometimes interactions happen outside of your measurable ecosystem that you'll NEVER have insight into. Executives HATE hearing this, but it's true.
I'd estimate that 20-30% of interactions with a company can't be tracked. This includes podcast subscribers, blog readers, someone who heard about your company from a friend who went to one of your events, and other individuals who haven't identified themselves.
Accurately assigning revenue to certain marketing activities like these is truly impossible.
2) Self-Identification
Many executives respond to the shadow funnel by suggesting the marketing team implement a self-identification mechanism that often asks something like this, "Where did you hear about us?"
The problem with this is that many users don't remember, or worse, will select something meaningless to proceed with whatever they're trying to do.
So you start from the top and begin doing several things like setting up a campaign framework, using UTMs, creating a list import template, and getting your house in order.
3) Lack of Campaign Framework
If you don't have a reliable and well-implemented campaign framework, many acquisition/attribution solutions simply won't work.
You NEED a strong set of campaign types, campaign member statuses, and orchestration to ensure that members are included in the programs, so that attribution systems can marry this information with opportunities created in association with the contacts that were part of those campaigns.
This is how revenue is attributed to marketing activities. No campaign framework, no attribution.
4) Poor UTM Coverage
Once you're set up with a campaign framework you turn your attention to ensuring your team is tracking UTMs consistently and effectively.
If you're an executive unfamiliar with UTM tracking, here is a nice post that delves into UTM Tracking in depth.
5) Missing Hidden Form Fields
If you've got your UTM tracking in place, that's fantastic for Google Analytics; BUT if you want to actually use those UTM values for acquisition/attribution tracking, then you need to get those fields added to ALL of your web forms.
With HubSpot this process can be done by marketing, but with other platforms (like Marketo) it is a little more difficult. It will require your development team to implement some code to add these UTM values dynamically to all forms, or you can painstakingly add every field to every form you’re using in Marketo.
I've run into situations where development can take months to implement this type of code, and that can seriously delay an attribution/acquisition tracking project.
6) Your Signup Page
More often than not, your marketing team will rely on the product team to make updates to the signup form for the application to properly track UTM values.
This will involve some development work that is certainly not trivial, and many product teams are heavily resistant to. They don't understand the purpose of doing this, they are reticent to work on it, and come up with many excuses for why things shouldn't be done a certain way.
In the end, the product team owns the application, and marketers often lack executive sponsorship to get changes made to the application to support marketing funnel tracking. This will create huge gaps in marketing reporting.
If you do, by some miracle, get product to implement UTM tracking effectively, you start to notice that UTMs have A LOT of shortcomings in the form of session management, cross-domain tracking, and other parameter stripping.
7) No Session Management
While UTMs will persist through sessions as it relates to Google Analytics, that does not mean that UTMs will persist through a session and populate the hidden form fields that many marketers rely on to capture this information.
All it takes is someone to click a link, visit a page, navigate away, and then back to that same page to ruin your measurements.
This is why many teams have resorted to storing UTMs in cookies or local storage to retrieve those values using a custom script.
8) No Cross-domain Tracking
Let's say someone comes to your site through a Google ad, clicks around your website, and then decides to sign up for your platform.
If they click a button to go to app.yourplatform.com from platform.com, your UTMs are not going to follow you by default. There is a good amount of work you need to do with your tag management solution, script configuration, and setting cookies in order to carry those UTM values over to the separate domain.
Again, this is something that you will likely need to work with development or the product team on implementing.
9) UTM Stripping
If worrying about session management and cross-domain tracking wasn't enough, many browsers, browser extensions, and email clients will now strip UTMs from links for privacy reasons.
This actively prevents you from measuring the impact of certain activities in your system, and there is nothing you can do about it other than creating your own unique tracking parameters to circumvent this stripping activity.
10) Domain Blind Spots
Finally, when it comes to UTMs, sometimes you simply can't measure something on a specific domain. It's not very common, but some SaaS platforms still don't allow you to include custom javascript or provide integrations with common tag management solutions.
In this case, you just are out of luck.
11) Lack of List Import Process
While trudging through all of the UTM madness, you’ll realize very quickly that not all touchpoints are digital, and your company probably exhibits at a number of events.
Getting leads from events into the system requires you to co-opt the UTM standard, create a list import process, and get people into your MAP / CRM in a very specific way.
If you don't do this then there is a huge gap in your acquisition/attribution data.
12) Shadow IT
Finally, once you start to feel like you've really got your house in order, you start to notice weird values in your system. Where could these be coming from??
Sometimes marketers work against themselves by using their own preferred tech instead of what a centralized MOPs team has sanctioned.
It could be a secret platform to manage social posts for a field marketer's specific region, an email tool to secretly send out newsletters in a specific language, or an app used at events to get people into the CRM.
There are tons of ways marketers knowingly or unknowingly work against themselves when it comes to acquisition/attribution tracking.
And with that fun realization, we enter the next painful phase of the attribution/acquisition tracking journey.
Phase 2: Data Cleaning / Maintenance Issues
Inevitably when you start reporting on the data that you've begun to collect your fellow executives are going to want to know why certain numbers aren't adding up in their minds (and in their various arcane spreadsheets), and a lot of this can be tied back to the topics in this next section.
13) Lack of Historical Attribution/Acquisition Data
Let's say that you've got 100K individuals in your DB. When you implement attribution/acquisition tracking protocols it is most likely that you'll NEVER know where these individuals actually came from or what their journey was like.
You can go back, cleanse that data, normalize it, consolidate campaigns/campaign members to fit your framework, and put in months of work to get everything perfect, but honestly.... the data was probably crap to begin with.
This is why it takes SO LONG to see results from acquisition/attribution projects. Depending on your sales cycle these measurements may not bear fruit for months, if not years.
I am a fan of having a hard cutoff, explaining that this portion of the database is a sunk cost, and only relying on data collected moving forward.
Not everyone is so lucky as to have this be the case.
14) Historical UTM Usage
Even with a future-looking perspective, you still have to contend with the past. You may have standardized all of your UTM combinations, built a tool to generate them, and enabled the entire team, but there will still be historical links pointing at your site that are tagged using old UTM patterns.
So, in addition to having your UTMs perfectly calibrated for the future, you have to set up protocols to cleanse bad UTM values that might make it into your MAP, and depending on how fast your processes run in the background this can be very tricky.
15) Editable Fields
Once your data has made the journey from MAP to CRM, and you're starting to feel good about how it's displaying you start running into issues of a different nature...
Employee error.
If you don't protect edit permissions on the acquisition/attribution fields in question your team (cough, salespeople) have free reign to mess that data up, and they will.
Your fields and data should be read only to everyone but the MAP integration user and admins.
16) Poor Contact / Opp / Account Associations
Many systems rely on an accurate picture of the contacts that are associated with accounts and opportunities being actively worked on to attribute correctly.
Marketers can only do so much when it comes to this. There are plenty of lead-to-account matching solutions out there, but in the end, it's the sales folks who control who is attached to an account/opportunity with specific roles.
You need an executive sponsor to lay down the law here on the sales side of the house. Or set up a validation rule or two to prevent opps and accounts from saving if they don't have any contacts attached 👀.
17) Bogus Product Data Model
As stated before... many systems rely on an accurate picture of the contacts that are associated with accounts and opportunities.
If you're in a low-touch, product-led company where buyers don't speak to a sales group, your product needs to be able to identify and sync to the CRM, the person who set up an account, the people who are using the platform, and many other account contact relationships.
If you're having trouble with this because your platform doesn't conform to the typical Account / Contact / Opportunity model that is pretty relied upon in the Salesforce ecosystem, well, good luck.
18) Poor Unification Rules
Companies oftentimes have tools running in the background like LeanData, Ringlead, Cloudingo, or ETL tools that attempt to dedupe or unify records. When this is done, sometimes certain fields are unified improperly.
As a simple example, let's say that we have 3 people in the CRM with the same email address and name. They are the same person but created at vastly different times. Each of them has a different lead source, acquisition program, and other information. When combined, the system for some reason is configured by an internal group (that isn't marketing) to use the values on the most recent record.
Well crap, you've just ruined your acquisition data.
Phase 3: Data Interpretation Issues
At the end of the road, after all of your attribution/acquisition data has been collected, cleaned, and reported on some in the executive suite will STILL not be happy with the results, and this can be caused by a number of interpretation-related issues.
19) Poor or Inaccurate Weighting
With systems like Full Circle Insights, attribution weights are determined BEFORE revenue is attributed to specific campaigns, and this weighting usually ends up getting determined by the person configuring the tool because no one else cares enough about the project to contribute.
The problem with this is that at the end of the day, various people might not believe the numbers that come out of the system. This inevitably leads to asking, "How they are determined?" That leads to trying to explain the various weightings, and when they hear that they weren't involved they refuse to believe any of the attribution data.
When you try to involve them, they are suddenly really busy.
20) Inability to Recalculate or Modify
If they aren't busy, and somehow you get the team to update the weightings, and maybe do extra work like updating all of the account/opportunity contact associations you've been yearning for... well, with some systems you don't have the ability to rerun your entire attribution. You're stuck with the data you have.
This is one reason I really like Full Circle Insights. They are built directly into SF, allow you to rerun models, and since they are built upon SF you can easily get data into the system that may have been missing.
21) Decay (And Other Common Cases)
Often a person will try out a platform, abandon it, and then another person from that same company will sign up and become a customer three years later. How do you attribute that?
Should the sale be associated with the first person's activity/information, or the second person's? Or both? How do you determine this?
Most marketers don't know how to answer these complex questions, and most acquisition/attribution systems aren't equipped to handle them from a technical standpoint.
Unfortunately, decay is just one example of a complex scenario that pops up.
22) Poor Executive Education
FINAAAALLLLYYYY, when your acquisition/attribution project inevitably comes to a close (it never actually does), then you have another unique challenge....
Executives are stupid... or that's at least how it feels.
You've educated yourself about all of the different multitouch attribution models like first touch, last touch, U-shaped, W-shaped, and blah blah blah, but there is no guarantee that the VP of Sales or your CEO has ever even heard of these terms.
Trying to set up exquisite attribution and acquisition tracking is great, but it needs to be understood that it will NEVER, EVER, EVER, EvErrrrrr be perfect. On top of this, it will most likely take a VERY LONG TIME to implement (9-12 months minimum).
Finally, even if you do reach the promised land, it may only tell you 70% of the story.
Here's the frustrating thing though... Even if you can get 70% accuracy, because it's not 100% it won't be trusted by many in the executive suite, especially if they have that one "trusted analytics person" that starts to poke holes in the data.
A bad or inexperienced executive team will default to rudimentary collection tactics sourced from the individuals who attacked you and have their very own specious collection tactics. The warble of these bureaucratic morons is "How many contact/demo forms did we get?!"
A seasoned executive team, most likely knows what is influencing revenue, and where customers are coming from, and will take the data you've collected with a grain of salt to more effectively inform and guide decision-making.
COLLECT THE DATA, TRUST YOUR GUT.
