When someone makes a purchase or signs up for your product, how do you know where they came from?
Especially when it comes to a top-of-funnel channel like social media, how can you ensure social gets credit for the impact it’s having?
Attribution to the rescue!
Seeing the impact of your social media marketing is definitely possible with an attribution strategy that assigns value to the channels that send new business your way. We’ll show you how to start in a lean way and then go deep with robust solutions as you scale.
Attribution can be a tricky science, so we’re hoping to give you some clear and actionable tips to start measuring attribution today. Whether you’re looking at running some omnichannel campaigns or if you’re wanting to measure the impact of some quick experiments, we hope these tips will have you covered.
Did you know …
The average consumer comes in contact with a brand 6 times on average before they make a purchase.
So how can you know which of those touchpoints deserves credit for the sale?
This is where attribution comes in.
Attribution is very close to another key social media metric: ROI. You probably hear about ROI more often because, well, to be honest, attribution is even trickier to measure than ROI — if you can believe that.
So maybe we should begin with defining each term and the differences between them.
The difference between ROI and attribution
ROI stands for return on investment. It is a dollars-and-cents measure of how much you earn based on how much you spend. There are related terms here too like ROAS, which stands for Return on Ad Spend. In either case, these are measuring your earnings.
Attribution is subtly different. Rather than measuring how much you earn, attribution assigns value to the channels that drove an outcome, whether it’s revenue-related or not.
For instance, ROI would measure revenue whereas attribution measures a purchase or a web visit or a download.
With attribution, you’re basically working backward from an event like a purchase or signup and figuring out what channel or channels get the credit for that event.
We’ll be focusing on attribution in this blog post, but it can get a bit muddled along with ROI. We’ll do our best to keep it all clear for you.
The three main types of marketing attribution
So attribution comes in a few different flavors. You have:
- Last-touch attribution
- First-touch attribution
- Multi-touch attribution
This diagram from digitalthought.me does a good job of showing these different models side-by-side. The multi-touch models include Even, Time Decay, Weighted, Algorithmic, and even more.
Here are some other resources that help explain these concepts because they can get complicated, fast.
- A Comprehensive Guide to Attribution
- The Complete List of Marketing Attribution Models
- What is Marketing Attribution?
Essentially, first-touch attribution gives full attribution credit to a person’s very first interaction with the brand, regardless of how many touchpoints occurred after the fact. If someone sees an ad on Monday and signs up after a Google search on Friday, then the Monday ad gets all the credit.
Last-touch attribution is the inverse. The very last touchpoint gets full credit. If you see an ad on Monday and sign up after a Google search on Friday, then Google gets full attribution.
We’ve found that last-touch attribution is the easiest to measure and is a great starting point if you’re just getting your feet wet with social media attribution.
However, these single-touch measurements won’t show you the complete customer journey like multi-touch attribution can. The only problem: multli-touch attribution is hard! and it requires a series set of marketing data ops.
For this reason, we won’t go too deep into multi-touch attribution — we’ll save the details for another episode — but just to give you a taste, here are a few of the models that you could consider with multi-touch:
- Linear, which measures all the touchpoints and assigns equal value to every one
- U-shaped, which gives 40 percent to the first touch and 40 percent to the last touch and then the remaining touchpoints share the final 20 percent
- Time decay, which gives more weight to touchpoints closer to the final conversion event and less weight to those at the very beginning of the journey
We’ll share some quick ideas and tools for multi-touch toward the end of the blog post, but for now, let’s transition from models into the specific tactics for measuring some of this.
We’ll start with a few of our favorite, scrappy ideas, some of which you might already be trying.
1. Track attribution with UTMs
UTMs are the extra parameters you can attach to links so that you can track your traffic from specific sources, mediums, and campaigns. If you’ve seen a funky long URL with question marks and ampersands in it, then you’ve probably seen a link with UTM parameters.
UTMs appear in your Google Analytics reports. You can see your traffic broken down by all the different UTM parameters you use.
So, let’s say you’re running a social media campaign about how dogs rule and cats drool. You can add a UTM campaign tag of “dogsrule” to all the links you share on social media. Then, when you hop into Google Analytics later, you can see the exact traffic impact of this campaign.
Yes, there will still be the referral traffic section in Google Analytics that counts your social media traffic in aggregate.
But with UTMs, you can also drill down into utm campaigns to see the total clicks from your specific “dogsrule” links.
If you have Goals set up in Google Analytics, you can then see the attribution to things like signups or downloads or purchases. If you have revenue numbers attached, you can even get the full ROI picture.
Same goes for tools like Mixpanel or Looker or any other analytics software that you use to track results.
Now with UTMs, you’re typically going to be getting last-touch attribution. Here’s how the tech works: The UTM parameter is added to the browser in a cookie. And if a person clicks on a link with a certain UTM then clicks on a second link with a different UTM, the second UTM is the one that will get credit. And of course, if someone clears their cookies or has cookies disabled, then the UTM attribution won’t work at all.
There are some tools that track the first click also. They call this the pancake stack — first click and last click. You’ll want to check with your analytics provider or your data team to see what exactly you’re tracking.
If you’re looking to add UTMs to your social media links, there are a bunch of easy ways to do this.
First, you can do it manually by typing in the parameters at the end of your URL. These are:
Additionally, we add these for you automatically with the Buffer product. On our paid plans you can customize the text here, too. So if you click on a Buffered link, you’ll be able to track exactly what kind of attribution it brought you.
2. Include a “How’d you hear about us” survey
These are going to be so much easier-sounding than UTMs. All you need for this form of attribution is a simple survey, placed after the event.
For example, we recently implemented this within our Buffer signup flow.
When you sign up for Buffer, we ask you a few questions, including, “How did you hear about us?”The options are:
- Friend, family, or work colleague
- Google search
- Online advertisement
- Blog post
- Social media post
With this data, you can then see exactly which sources send you the most attribution. We tend to think of “how did you hear about us” as first-touch attribution, though it does remain a bit open to interpretation.
At Buffer, social media accounts for our third-biggest attribution source at 10% of signups. Our top two are Google searches and word of mouth.
Generally-speaking, you could expect a certain margin of error in the survey results from people giving random answers or clicking incorrectly. Still, it’s useful data to know!
3. Try bit.ly URLs to track attribution
These can help measure ballpark attribution numbers, especially when it comes to the middle of the customer journey.
Let’s say you’re wanting to see how many people came to your site from the link in your Instagram bio. Well, you can hide UTM parameters within a shortened link, and then go to your bitly dashboard to see some robust stats about clicks and timing.
This can become especially useful when you’re using a single bit.ly link across multiple channels because bit.ly will give you the conversion by channel. It’s a real-time view of your stats, versus a bigger dashboard like Google Analytics that excels at looking at previous days and macro views.
We’ve touched on some quick ways to get started with attribution — adding UTMs to your links, putting surveys onto your website, or using bit.ly to track click performance. These are great for single-touch moments of attribution.
4. Multi-touch attribution resources
Resource #1: Find a tool that will help you track the data and then place it into a multi-touch model for you.
Here’s a few ideas of what tools to try:
Bizible — This tool specializes in end-to-end marketing attribution. It was acquired by Marketo last year, so it has a ton of resources behind it. It’s one of the biggest and best tools out there for attribution. It’s not cheap — you need to hop on a demo call to get pricing. But it is an awesome solution especially if you’re wanting to integrate with a CRM like Salesforce.
In addition to Bizible, there are also tools like Impact and the appropriately-named Attribution.
Also, some of the major networks like Facebook and Google have their own attribution products that can help with multitouch attribution. Facebook Attribution covers all the different touchpoints that your user might go through in the Facebook network, which includes organic and paid on Facebook and Instagram, plus any placements on Facebook’s Audience Network and Messenger. Google’s attribution solution is now built right into Google Analytics so you can build your attribution model right in there.
The other way to look at multi-touch modeling is to be consistent with the way you model and track your social ads. This can be a huge component of the marketing journey for some brands, so it’s important to get these settings right.
Specifically, you’ll want to look into the attribution windows for your ads, and you can do this right in the ads managers for all the social networks.
Generally-speaking, there are three types of attribution windows for these ads: engagement windows, click windows and view windows. This means that if you were to set, say, a 7-day click attribution window, then the ad would take credit for the conversion for anyone who converted within seven days after clicking the ad.
There are a lot of different ways to set up these windows. Pinterest alone has 24 different options for you. The key is that you want to be comparing apples to apples so that you can easily compare the effectiveness of each channel.
One of the best places to start is with a 7-day click attribution window. This can be the same across all your social channels. You can start here and then tweak and adjust the window over time once you start seeing the results.
Of course, it goes without saying: In order to get value from attribution, it’s important to know what your social media goals are.
Since attribution is based on the events and behaviors that you want to drive … then you need to know what those events and behaviors are!
We’ve seen a ton of great examples of the different ways that brands are measuring their social media impact. When you have revenue and attribution modeling in place, you can find some really cool insights.
Here are just a few stories that eConsultancy shared in a recent blog post:
- KLM looked at sales generated from customers who visited their site directly from social media and found that these customers contributed 25 million euros in sales.
- MADE.COM compared the average order value of people who went to their site from social and their site average. Their social media fans spend, on average, 4 percent more per order.
- Incontact trained half of their sales reps on social selling and found that these sales rep had a 160 percent bigger sales pipeline and achieved 215 percent more revenue than those without.
It’s important to remember that when you use social media as an engagement and brand-building channel, you might not generate many leads or sales directly from social media. But you would indirectly.
We’re hopeful that these attribution ideas might give you a sense of how well social media impacts those indirect leads and sales and that setting up attribution can help you know your channels better.
How to say hello to us
We would all love to say hello to you on social media – especially Twitter!
Enjoy the show? It’d mean the world to us if you’d be up for giving us a rating and review on iTunes!
About The Science of Social Media podcast
The Science of Social Media is your weekly sandbox for social media stories, insights, experimentation, and inspiration. Every Monday (and sometimes more) we share the most cutting-edge social media marketing tactics from brands and influencers in every industry. If you’re a social media team of one, business owner, marketer, or someone simply interested in social media marketing, you’re sure to find something useful in each and every episode. It’s our hope that you’ll join our 27,000+ weekly iTunes listeners and rock your social media channels as a result!