Keep track of goals using the Linearity Indicator in Adobe’s Analysis Workspace

There is an universal truth in life: Inspiration always strikes when and where you least expect it. The same happened to me the other day, when I was reading High Output Management by former Intel CEO Andrew Grove. While the book is definitely worth reading for anyone interested in management, analysts can benefit just as much from reading it to get inspiration for valuable performance indicators and visualizations. Quite early in the book Grove presents one of his favorite visualizations to track progress towards specific goals: The linearity indicator. This chart shows the current progress towards a set target and where the performance might be heading. Here is his example for a hiring target from the book:

My initial reaction was “wow, this is super cool and simple to understand”. If the current progress is above the linear progress, we’re in good shape to reach our goals. If it is below the line, we need to pick up some speed. Super easy! My second reaction (naturally) was “hey, I can build that in Analysis Workspace!” All we need are a few calculated metrics and a line chart. So, let’s get our hands dirty, start building, and see how far we can get!

Step 1: Preparing Analysis Workspace

To start, we should define how we want to keep track of our goal, starting with what we actually want to keep track of. I’ll pick the example of online orders for this post, but any other metric would work just as well. We can drag that metric into workspace to get started and add a line chart as pictured below:

In my case, Workspace already chose the correct date dimension for me. We will keep track of our monthly online orders here, but the time granularity could just as easily be changed too. As an added bonus, we will build our metrics in a way that allows us to change the resolution on the fly! So if later we decide to keep track of our weekly or daily online orders instead, we just need to change the dimension. Super neat!

Step 2: Adding a goal and linear target to Analysis Workspace

Now onto our first Calculated Metric. We need to have the constant line on top of the chart, showing the total goal for the current reporting period. Adding such a static number is super easy, we just need to click on the little “Add” button on the canvas:

This will give us an empty box to fill in our goal for the period. In my case, I’ve set a random goal and named the metric “Yearly Online Orders Goal”. When I add it to the Freeform Table in Workspace, it shows up like below. Note how the value is the same for each and every month, which is exactly what we want!

Neat! To get to our linear goal progress indicator, we need to create another Calculated Metric. This one is a little more complicated. Here is what we are going to do: We need to divide the total goal we created before by the number of months (or weeks, days, etc.) we have in our table. That average per time unit then needs to be summed up for each row. Speaking Calculated Metrics, we need to use the Cumulative function to add together the rows in our table. The value is the goal divided by the number of rows in our table (using the Row Count function) as shown below:

I’ve named my metric Linear Online Orders Goal and added it to the table, where we now see how our orders would add up if we had exactly the same number of orders each month:

That looks great already! Now we can see how we would need to perform if we wanted to achieve our yearly goal in a world without seasonality. Now let’s get our actual performance in there too!

Step 3: Creating a cumulative performance indicator in Analysis Workspace

While the monthly performance is already in our table, we need to display it as the cumulative of the previous months. Easier put: Each row should contain the sum of the online orders for the whole year until that point. This can again be achieved using the Cumulative function in the Calculated Metric builder. We could just use that function on its own, but I’ve wrapped it in a logical test that only returns a value if that month actually had orders. You will see why we want this in a second, but here is the metric so far:

Adding this new metric looks pretty good already on our Line Chart:

You can see that, thanks to the IF-condition above, our new metric goes to zero as soon as there are no more Online Orders. But why would we want that? Because we have this nifty option called “Interpret zero as no value” in the column options of our Freeform Table:

Upon activating this option, you will see that all the zero rows disappear from the Freeform Table. But that’s not all! This option also affects the Line Chart, where our metric lines now end where the data ends:

Awesome! We could already offer this chart as-is to our stakeholders! With such a cool visualization, it would be easy for our sales managers to track their achievements and see where the trend is heading. But if you know me, you know that I like to take my metrics to the next level, so let’s get predictive!

Step 4: Adding a remaining orders metric to Analysis Workspace

If you take another look at the photo from High Output Management at the beginning of this post, you will notice that we have to add two more lines, labeled “Will me make it?” and “Miss it?” in the original. Let’s start with the “Will we make it?” line first.

To create a Calculated Metric for this line, we first need to think of what we want to see. For my chart, I want to make a direct connection from the end of the cumulative performance line from step 3 towards the actual yearly goal. To do that, we need to find out how many orders are left until we reach our goal and distribute the yearly remainder over the remaining months. Let’s do it step by step.

To get the number of remaining orders, we first sum up how many we have already. This can be done super quick: Just simply right-click the Online Orders column and, in the Create metric from selection menu, select “Column sum”. This will give you a metric that has the sum of the column in each and every row:

Nice! Now, we do the same right-click magic with the Yearly Online Orders Goal column and the Column sum Online Orders column, but select Subtract as operator, giving us the remaining Online Orders for the year:

To get to our monthly remaining orders, we now need to find out how many months are left in our date range. This is a prime use case for a Conditional Counter metric, which I’ve explained in a dedicated post already. I’m not going to cover the full process here, but how it works is quite simple: Create a Calculated Metric that returns 1 if that month had any online orders, then sum up that column. The formula for the first part could look like this:

The sum is again only one right-click away and shows us how many months have already passed:

Now we just subtract the total number of rows from the number of months that have already passed to get the number of remaining months:

Nice! Now that we know how many orders are left for the year and how many months remain, we just divide them again to get the monthly remaining online orders:

We are almost done with this step, I promise! Now that we have the remaining monthly orders, we do the same cumulative calculation as before and only return it on rows without any online orders. Then we just have to add them to the actual cumulative online orders and we’re done! Here are my intermediary columns and the end result:

The resulting Calculated Metric is quite complex by now:

That’s a lot of metric! But the result looks super cool in our line chart. You can see how the green line stops at one point , revealing the orange line behind it that ends exactly at our yearly goal:

We’ll clean things up a bit at the end of our experiment. For now, we have one more line to add!

Step 5: Predicting order performance in Analysis Workspace

To add the last remaining line to our chart, I’ve opted for a simple linear indicator: We’re going to calculate the average monthly orders for the year until now, then draw a line into the future that continues this trend. The average is super quickly created using the Mean function in the metric builder. Make sure to not check the “include zeros” option so we don’t count the empty, future rows. The table with the metric should look like this:

Now that we have those, we can do the same steps as before: Create a conditional mean online orders metric, add it to the actual online orders, and wrap everything nicely in a Cumulative metric:

Nice! Now we just need to clean our table up a bit to make everything look a bit nicer.

Step 6: Finishing touches

In our final table, we can remove all columns except the Goal, the linear monthly indicator, the linear forecast, the on-target forecast, and the actual performance. My table now looks like this:

The best thing about this table: We can hide it and rename the columns in the line chart, so nobody needs to know about my lack of naming things properly! Based on the table shown above, this is how I renamed my lines on the chart. I also added a little note about how to read the chart:

How cool is that looking! If you’ve followed along, you can now switch the granularity to daily, weekly, or quarterly intervals and the chart just automagically adapts! You’ll also notice how the order of metrics in the table influences which line overlays which other lines in the chart, so make sure you put the current performance into the right-most column.

Wrap up

I hope you enjoyed reading this post as much as I enjoyed writing it and creating all those nice Calculated Metrics. For me, this is another great example of how versatile and flexible Analysis Workspace and Calculated Metrics can be, if you just take the time to learn about all the options. My favorite detail here is how the Line Chart respects the settings to treat zero as no value from the Freeform Table.

Now that you have read this post, I’m curious to hear what use you may make of it. I can think of a bunch of applications myself, both inside and outside of sales teams, but hearing about what you would do with this tool would really make my day. Until next time!

Frederik Werner

German Analyst and Data Scientist working in and writing about (Web) Analytics and Online Marketing Tech.

2021 – current Adobe Analytics Champion

EMEA Adobe Analytics User Group Lead

Adobe Analytics Community Advisor

www.frederikwerner.de/