Twitter recently rolled out App Install Ads worldwide including India. While earlier also it was possible to drive App Install campaigns using Twitter Cards, they have now made it much easier removing the need to add meta tags to your website.
I decided to try out Twitter App Install Ads for its effectiveness. Here is my quick analysis starting from setting it up to end performance:
1. Setting Up — Adding new app is pretty easy — just enter your app IDs and it will pick up all the attributes from App Store
2. Select Targeting Parameters — There are 4 ways to target your audience. (Television targeting is available only in select few countries)
3. Define your budget and CPC bid — CPC bid can range from $0.01 to $2.00
4. Integrate conversion tracking — Twitter has tied up MAT, AppsFlyer, Fiksu, Kochava & Grab Analytics as measurement partners. You have to open account with one of these to track conversions and engagement. I hope Twitter introduces some callback URL option in future to allow any measurement tool to track engagement.
5. Go Live — Just create a new tweet, add image and press save to go live. Twitter recommends multiple tweets for better conversion. I tried with just one.
So does it work?
1. Volume — Despite keeping liberal targeting & budget, Ads could not generate much volume. After 1 day, it could not even exhaust the small budget I had set up for experimentation. Facebook with similar targeting would have gobbled up every single penny in my account within few hours
2. Cost — Twitter gives you ECPAC (Effective cost per ad-click) which is effectively CPC. Twitter surprisingly delivered a click at half the price of FB
3. Conversions — This is where things start going down for Twitter. FB generates 5X better CTRs. Click to Install ratio aggravates the case further. FB converted 40% of those clicks into installs compared to 20% by Twitter.
4. Engagement — Its even steven here. ARPU of Twitter Install is similar to FB’s. Too early to create cohorts & LTV but ARPU is a decent indicator for knowing the quality of a user.
Wait — Dont make any conclusions yet. Point no.3 above must have suggested that Twitter isn’t a great option. Now see the table below to visualize the whole conversion metrics (Note: Dummy CPC & Impression data used for simplicity)
Check the last column — CPI delivered is identical. So the only downer for Twitter app install campaign is its poor volume. I reckon that conversion for Twitter would improve if they have more inventory to target. That will make Twitter a worthy competitor to FB in India/elsewhere for app install campaigns.
Important Note: Above results are for a lifestyle app appealing to wide audiences. Result for each app may vary depending on its popularity, TG , category and many other parameters. Sole purpose of this analysis was to give you an overview on how to measure your campaigns on Twitter.