When you create an ad on Twitter – and I’m talking only about Promoted Tweets – you pay for an “engagement”. This can mean a number of widely different things, as defined by Twitter.
screenshot from Twitter’s ad CMS
Twitter will charge you for retweets, follows, any Twitter event, as well as a click anywhere on your tweet whether it’s a click on your username or on the tweet or a link in the tweet. And they will charge you the same amount regardless of which action your audience is taking on that tweet.
Is this good or bad?
What if everything at Target was priced the same: the average of the prices of everything they had?
It would suck when you had to buy detergent (you’d go elsewhere), but it would be great if you wanted a washing machine. And if you had an entire shopping list, it would boggle your mind planning your shopping. The metaphor would be more applicable if you couldn’t choose what you wanted to buy – you could just set a total budget and be given a few random items. Not a great deal, is it?
There are 3 things an advertiser might want from a Promoted Tweet:
- Direct Response: make a sale or generate a lead, immediately.
- Virality: Seed some content with traffic in the hope of it going viral (which in itself is a means to some other end).
- Followers: Get followers to be able to make a sale to them in the future.
An advertiser’s goal for one ad may be more than one of these, but very rarely will they be worth the same to the advertiser.
For a direct response ad with a link to a product page, followers or virality is not important because the goal is to sell the product. In fact, it is better for the ad to be precisely targeted at users likely to buy the product. It being retweeted to the Twitterverse will bring in a lot of traffic, but largely irrelevant traffic.
If the goal is to make some content go viral – to perhaps generate ad revenue from the content (e.g. YouTube videos) or for some branding – and to create some seed traffic to set off the virality, clicks on a link in the tweet will be worth less than retweets.
Getting followers is always valuable because you can reach out to them for free at any time, although not in a targeted way. How much an advertiser would want to pay for followers varies often.
Lastly, if a user clicks on the tweet username, that counts as an engagement too. I’d consider this nothing more than an ad impression, and certainly not an engagement. Charging the same $ for this as for a follow or a retweet is almost ridiculous.
Optimizing RoI is easier when you know the R and the I.
If the ad system knows exactly whether the advertiser wants clicks, retweets or follows, and how much the advertiser is willing to pay for each, it can optimize ad delivery accordingly. If the primary goal is to get retweets, ad delivery can be tuned to reduce exposure to user segments that are clicking but not retweeting. That audience could then be shown ads that need clicks and monetized better.
Setting multiple goals lead to complex trade-offs.
With multiple desirable actions, advertisers need to now place different bids for different actions. It is not simple to work out how much more a click is worth than a retweet on a direct response ad. Or the converse when virality is the goal. Also, at times the ad delivery system will face complex tradeoffs when allocating campaign budget between a secondary goal (e.g. follow on a direct response ad) that is achievable now, or waiting for a non-deterministic future event for the primary goal (click).
The solution shouldn’t be to avoid this complexity, but help the advertiser navigate it. Instead of bundling every action into an “engagement”, a good starting point could be to at least let the advertiser specify which action is their primary goal.