Improving on Data Visualization - Showing Change in Cost Per Engagement Based On Promotion Duration.
Comparing data visualization methods such as bar chart and area chart to show change.
Back in 2015, I did a blog post on how to reduce the cost per engagement or views in Facebook Ads.
The essential idea [not mine, had read somewhere] was to treat it as a demand / supply equation. The two scenarios are : If your demand is $100 worth of ad impressions over 1 day compared to $100 worth of ad impressions over 4 days, you're sending different signals to Facebook.
To use an analogy that married men can relate to...
If your anniversary is 24 hours away and you're scrambling to find that perfect gift, there's a good chance that the shopkeeper will sense your [totally not sweating] body language and NOT bargain much on something that you decide to pick up!....Compare that [possibly real life] scenario to you just looking around for gifts a week [or month] in advance...you won't be panicking, you'll take your time in checking out different options and maybe even being ready to walk away from a deal.
Right, going back to FB Ads... you can download ad sets and then know the number of days it was promoted for and the avg. cost per engagement for it. This is the visualization I had in mind [dummy data here...try out with your own ad account]
Bar chart - Showing cost per engagement by promotion duration
Ok, so we can see that promoting Facebook posts for 1 day is the most expensive and at some point, it starts levelling out.
Here's a spin on the same data - with better effect [I think]
Area chart - Showing reduction in cost per engagement by promotion duration
The cost per engagement is instead shown as the decrease in cost per engagement, compared to promotions that lasted 1 day only. So [CPE for 2 days - CPE for 1 day]/CPE for 1 Day.... [CPE for 3 days - CPE for 1 Day]/CPE for 1 Day...and so on.
With this area chart, we can know that promotion posts for 4 days is 50% cheaper than promoting content for 1 day [assuming same ad optimization objective, audience demographics].
Meaning, you could be getting 2X the number of engagements from the same budget - if the demand side of the equation was cooled from 1 day to 4 days. You can also see that you get diminishing reduction in CPE/CPV as the days progress...bottoming out at 57% for day 6....content freshness would probably play a big role in this...it could even happen that the inverse starts happening and you see an upward movement at day 6 / 7...so an increase in CPE/CPV.
Both cases use the same data but the latter one creates a more dramatic effect and quickly puts things in percentage [which helps everyone understand stuff without going into the actual CPE down to two decimals...]
What are your thoughts on the two visualizations? How would you improve them?