Monday, September 7, 2020

Facebook Attribution Training

The path to purchase is often fraught with danger for the digital marketer. When you’re creating a purchase funnel it’s easy to forget that it’s PEOPLE in the funnel (not just impressions, video views or conversions) and a lot of the time, people don’t make any sense. They see an ad on Facebook and of course they immediately click to your website and purchase! Isn’t that the dream…but we all know that isn’t exactly how it happens.

Now more than ever customers are researching, watching, waiting. They know when the sales are, they know that brands will get them back with a discount shown in an ad. They are going to multiple places to look at products, comparing pricing, shipping and other value adds, ALL with different accounts (Facebook, Google and so on…) So... how on earth do we track this? There is no direct path anymore and there probably never was but now we know it. 

This is what makes attribution such a touchy subject.

There are so many touch points in the customer journey, it’s often complicated to attribute your sale to one action. Just like you couldn’t make a goal in soccer without an entire team, it’s hard to give all the credit to one action that a customer takes, it is often a combined effort. 

So, what do we do? How does an agency show their clients that they’re not only doing their job, but they are in fact KILLING it? You need to have data. Show us the money! This is where attribution metrics come into play. Facebook is now making it easier to track and attribute your sales to the marketing efforts that you’re engaged in.

Firstly, there is no perfect model. There is no way that you can attribute ALL your sales from ALL your ads, there will always be a margin of error. There are a few models that we’re currently using as the industry standards, they vary from first click, to last click, to position based. All of these rely on rules, an ‘If THIS then THAT’ idea. 

If a user takes THIS action (clicking on your ad) and then reaches THAT part of the process (they make a purchase), that is classed as a conversion. BOOM! Easy. 

But what if a user clicks on your ad on their phone, gets distracted by a text, forgets about your product, goes to work, then that night, goes to Google on their laptop, searches for your product, then clicks on a shopping ad in Google and buys. Who gets the credit? The first ad they saw? The last ad they saw? This is where you need a smarter model. 

Statistical attribution uses a combination of things like machine learning, algorithmic learning, and big data to track and then distribute the credit for a sale to each step of the journey. Think of it like participation awards for each part of your funnel. Each part of the journey will get credit where it’s due. 

This is not really a new way of looking at your data. Marketers have wanted to be able to take the glory for sales for the longest time! 

We’re now able to see the metrics with far more accuracy than ever before. No more spray and pray with your advertising. Facebook has made it easier for us as marketers to see with increased accuracy, what is working best for a business in direct relation to your Facebook ads and other marketing campaigns.

It means that we will be able to optimiZe your campaigns with accuracy and stop wasting money on inefficient campaigns that might have skewed metrics. We will be able to better use our existing spend to generate MORE sales and customers. Which at the end of the day is what businesses want, more customers for the same amount of money.