A consumer notices an ad in a mall, sees it again later on Twitter, a few days later he remembers about it and visits the website, and finally and makes the purchase there. Who gets the credit for this conversion?
Marketing attribution is the art of assigning the right amount of credit to all the touch points involved in the sales cycle. In this day and age, the internet is bubbling with the rise of new channels to reach out to your customer, and everyone has multiple devices to access these channels on the internet.
This coin has two sides. On one side, you think “Well, marketers must be happy. Plenty of digital channels means plenty of chances to attract customers. Which means more conversions”
But here’s the flip side: The presence of multiple channels may help increase conversions, but it also makes it increasingly difficult for marketers find out which touch points are doing well, and where to invest their marketing budget in.
Multi Channel Attribution is hard work – from a plethora of online and offline marketing channels, it is no mean feat to identify exactly which touch point led to a conversion. You could say that, at this point of time, it is impossible to identify which touch point contributed how much to the conversion.
But why bother with Multi Channel Attribution in the first place? If you’re getting returns, people are converting left, right and center, why drill down?
Two reasons. First of all, to give credit to someone/something when it’s due, and secondly, to optimise marketing spend – basically to invest more on the marketing campaigns that work and utilize the budget more effectively.
A perfect multi-channel attribution model is the holy grail of all marketers. It is extremely cumbersome for marketers to build even a partially decent attribution model. In fact, it is so frustrating for marketers to deal with the variety of ways in which we interact with their marketing strategies, that they start to wish the worst for our online lives!
Imagine this scenario: A marketer struggling to build the perfect attribution model meets a genie who grants him three wishes. What do you think he would ask?
Marketer Wish #1. I wish my customers could afford only one smart device
As long as the consumer uses the same browser on the same computer, digital marketers can track their click-paths for as long as they want. They do this using cookies – tiny text files placed on your computer by the websites that you visit.
But here’s the deal. We are the four-screen generation – TV, Desk/Laptop, smartphone and tablet.
The average consumer sees an ad on TV, Googles it on her smartphone on the way to work. Later at night after work, she visits the website directly on her laptop, and makes the purchase.
So what is the aim of multi-channel attribution here?
To rightly apportion credit to the TV ad, the mobile ad, and also the website.
As you might have guessed, it is extremely difficult, even impossible, to track the journey of one user across devices. Because consumers can switch between devices – cookies cannot.
If the genie did grant this wish, the marketer would get a complete idea about the effects of his campaigns across different devices, and which ones work best. But the reality is that, this is never going to be true.
It is increasingly difficult – almost impossible – for a marketer to understand how their advertising and marketing efforts reach consumers across multiple devices. Targeting consumers and collecting their data across multiple devices are rising concerns among marketers today.
Owing to these concerns, tech-giant Google itself has now come to the scene with a new feature in Google AdWords – ‘Estimated Cross-Device Conversions’. Using anonymous, aggregated data from previously signed-in users, Google calculates an estimate of the total number of conversions, including conversions that happened across devices – it helped to identify 16% more conversions.
However, this feature is still in its conception – in 2013, it was introduced for AdWords, and in 2015 it was rolled out to DoubleClick advertiser products.
Marketer Wish #2. I wish my customers were forced to make online purchases in one step.
If this were true, one would see an ad, click it, and purchase it rightaway.
Ad –> Purchase
Here, there’s only one step in total, so he assigns all credit for this conversion to the ad. He is happy and sleeps peacefully.
Wait, wut? The average consumer doesn’t blindly purchase – she sees something on twitter, directly visits to the website to learn more about the business, did a Google search, clicked through a referral, again went to the website and finally made a purchase.
Look at the sheer number of digital channels through which she has traveled with regard to this purchase. Tools such as Google Analytics help marketers see the complete click-path of a consumer, which looks something like this:
Twitter –> Direct –> Organic Search –> Referral –> Direct –> Purchase
So how does the marketer apportion credit?
Web analytics tools use many attribution models: A common one is the Last Click Attribution Model, which assigns all credit to the last step just before conversion. In the case mentioned above, if the marketer applies the Last-Click model, the final direct search would get all the credit for this purchase.
What?! That is unfair – What about the social media channel that introduced her to the brand in the first place? What about the referral – it probably steered this conversion the most.
So, no surprises there, any attribution model that puts all its weight on a single touch point is flawed. The idea is to do exactly the opposite – assign weights to all touch points or the view-through, so that the model is balanced. This, again is more of guesswork than perfection. Trying to build the perfect attribution model, even if the different channels are all online AND on a single device, is a wild goose chase.
Multi-Channel funnels in Google Analytics also help a great degree in assigning credit and getting insights into what drove each conversion.
Yes, it is possible to track the click-path of a consumer who uses the same device on the same browser. But finding an attribution model or multi-channel report that has the perfect algorithm to apportion right amount of credit to the right touch point? Next to impossible.
Marketer Wish #3. I wish my customers were socially inept beings who loved the comfort of their homes and despised human interaction
Let’s take things one step further and give marketers their worst nightmare – consumers not only browse on multiple devices and go through multiple websites, they now want to talk to the store on their phones. Even worse, they want to physically visit the stores! *Gasp*
As we discussed earlier in this post, tracking web traffic on a single device/multiple devices is difficult in itself. Tracking foot traffic is even more cumbersome. And mind you, foot traffic is increasing with each passing day.
Contrary to what this distressed marketer wishes, most humans like human interaction – especially when it comes to purchasing things using their hard-earned money. Consumers are switching between the online and the offline spaces so frequently that the line between the two worlds are blurring. They now live in a ‘non-line’ world.
The strongest seed for the non-line world is this: Calls. The smartphone revolution has ensured that anyone who wants to make a call today does it easily without much hassle.
Websites have click-to-call buttons that facilitate instant calls. Google search results in your smartphone display a call button that takes the consumer straight to the dialpad with the number on it, all ready to dial – how convenient is that?! Also, call rates have gone down and people do not look at phone calls as a luxury anymore.
Performing attribution in the non-line world is not easy. Tracking online marketing to offline conversions is not easy, because there is no exact way to identify which online touchpoint prompted a consumer to make an offline purchase. However, there are a few workarounds:
- Send unique coupons through e-mail and social media, prompting consumers to use it in the store
- Conduct controlled experiments – Take two sets of people, with all the variables same but one. Perform tests on these sets, analyse the results from both sets.
- Call Tracking – Track all calls coming in to business. Assign a unique phone number for each channel, which is not available on any other channel.
The perfect attribution model does not exist.
Multi Channel Attribution is complicated, and is still in its genesis. The three pain points described in this post will continue to bother marketers for a while, and the best solution is to forget perfection, and work on what’s optimal for the business. Excessive nitpicking on your attribution model is a fool’s errand.
And even when you decide on the perfect working attribution model that gets you enough ROI for your marketing spend – don’t stop there. Attribution merely assigns a value to a touch point – the responsibility of increasing the value of each touch point still lies on you!