Even when you’re told it’s not possible.
I recently helped a client with an issue that’s far more common that I’d like to think in modern web development. They were paying for an e-commerce hosting and processing service that was unable to pass Conversion data into Google Analytics. They have the data, but the way the service is configured, there’s no way to push the data into a client’s analytics account. Without the ability to see what’s converting, and where those conversions are coming from, it becomes difficult to build a well-functioning marketing campaign. With no tracking, they may as well be putting their money into radio and tv ads.
So, how do we solve this problem? Short of demanding that the developers build systems that take this type of need into account, we have to get creative with the data we do have and learn to live with ‘best estimates’ and averages rather than exact data.
Step 1: Determine what you already have
One of the things that the service provides is a daily and monthly sales summary. In that summary, we know how many sales were made, how many items were sold, what those items were, and the total value of the sale. For most business owners, this is seen as enough information to keep going. But what if you want to know where your sales are coming from? Are they coming from search, Facebook, AdWords, somewhere else? With this system there’s no way to know. We need more.
Step 2: Determine your averages
The first step is to determine how you want to start counting your sales. The easiest solution is to get the value of an average sale. You can do this by dividing your total dollars in sales by the number of sales. If you have $12,000 in sales in one month, and a total of 341 sales, this makes your average sale $35.19.
$12,000 / 341 = $35.19
Now we have our average sales value, but we still don’t know where these sales are coming from.
Step 3: Using Add To Cart and Google Tag Manager
One of the areas that we did have access to was the Add to Cart button. Since this happened before the actual shopping cart we were able to have our own Google Tag Manager code on that page. We created a Trigger and Tag designed to do nothing but count the number of Add to Cart clicks we had.
Setting this up is straightforward. In GTM add a Trigger called Add To Cart. Select ‘Click – All Elements’ under Trigger Type and then select a click element that is unique to the add to cart button. In our case, it was the ‘Click Text – add to cart’.
We then created a Tag called Add to Cart as an Event in Google Analytics. We add a Category called Click, Action called Add to Cart, and a Label called Add to Cart. We will come back and fill in the value later.
Step 4: We wait
This part is the hardest. We wait. We need to let the Add To Cart Event run long enough to gather enough data to be statistically relevant. Ideally, I would let this run for a full month to find out just how many Add to Cart events happen in a specific month. In this case, we had 423 Add To Cart events compared to the 341 completed transactions. This tells us that we had 83 Abandoned Carts in that month. There will be some of these that came back later and finished their transaction, but this is just one element of why we use ‘Fuzzy Math’.
Using these numbers we can calculate our Completed Cart Rate.
( 341 / 423 ) * 100 = 80.61% Cart Completion
Step 5: More Fuzzy Math
Now that we have our Average Cart Value and our Cart Completion Rate, we can determine our event value for each Add to Cart click. We do this by multiplying the average cart by the cart completion rate.
$35.19 * .8061 = $28.37 Add to Cart Value
Now that we have our Add To Cart Value we can go back into Google Tag Manager and add $28.37 to the Value section of the Add To Cart Tag. Now, every time someone clicks Add To Cart an event is triggered in Google Analytics called Add To Cart with a Value of $28.37.
Step 6: Wrapping it all up
Finally, we can go into Google Analytics and check on the Events. Go to Behavior, then Events, and click Overview. You can look at the category level or the Label or Action level, I prefer the Action level. Now we need to select the Secondary Dimension -> Acquisition -> Source Medium to find out where all of these Add To Cart actions came from.
Is this 100% accurate? No. Is it better than not having any data at all? No doubt. It would be great to know if your Facebook traffic has a higher or lower average cart when compared to AdWords or Search. But that isn’t possible in this scenario. Some ways to get around this are to create unique landing/sales pages for each traffic funnel or just run one type of paid traffic at a time to see if there’s a difference.
The key is to learn to relax your data requirements and learn how to use what you have rather than stress over what you don’t have.
