I recently bought Twitter ads and used their new lead generation “cards,” a type of ad where Twitter users can subscribe to your newsletter without leaving Twitter. Here’s what my experience was like.
Background: Twitter Lead Generation Cards Sponsored Tweets
These ads have drawn a lot of attention, because acquiring subscribers is always a challenge for newsletter publishers and these cards promise to help them do it easily and at scale. Provided you can monetize the subscribers, who wouldn’t want to buy these ads?
I personally was pretty excited by the prospect, because I had used ad.ly in the past to buy tweets for The Advanced SEO Book, as part of my list building for that. And they worked really darn well. As with other ads, they required optimization, but I saw out of the gate that I got leads – lots of them. (I stopped using ad.ly when they pivoted from being a mass market for ads to being a celebrity endorsement platform. I don’t have $25,000 for a tweet, thanks.)
I run RockThePitch.com, a soccer training newsletter.
There was a bug getting started, at which point their support kindly helped me.
Here’s the first sponsored tweet I made:
And here’s the second ad:
Note that in people’s timelines, they don’t see the link but instead see the lead gen card as shown below:
Seems pretty straightforward:
- Attractive offer? Check
- Clear language? Check
- Call to action? Check
The areas where it’s a bit weak are that:
- the shoes don’t contrast with the call to action but instead blend in, and
- there’s a slight disconnect between the headline, which just asks if people want the cleats at a discount, and the call to action which says get the coupon by email.
Here’s how the first tweet performed along with the above lead gen card. (I used the same lead gen card for both tweets. though one was a copy of the other with different backend tracking. This changed nothing to the user experience though, so doesn’t affect conversion.)
The problem as you can see is that I’m paying a lot for engagement – here only clicks – but the meaning of these clicks includes everything from clicks on my username to clicks on the photo it seems. Did they at least pay off in new followers? Nope. Those were really worthless clicks.
Some showed up on my site, but Google Analytics shows dramatically fewer (12-20 max, far from the 75 listed total for this and the other card I’ll show in a moment.)
For now, Twitter is optimizing to increase engagement (e.g. showing the ad to those most likely to engage), without regard for conversions. It makes sense from their perspective, but I kept racking up engagement but there weren’t incremental email signups. That doesn’t make me want to buy more, obviously.
This isn’t new of course. Facebook sells the same nonsense to its advertisers, telling them to buy engagement. You got a like, 3 shares and a clappy rabbit! Woohooo!
The fact that others do something doesn’t make it right, however. I hope that Twitter allows its advertisers to buy what they want, and not bundle together brand (“engagement”) with direct response (“clicks to my website or leads”). It would make smart business sense as well, leading to higher bids on the part of its advertisers, who could show better ROI.
My second sponsored tweet with the above card design was a bit more specific, as I mentioned. I called out Robin van Persie fans, since I was targeting users similar to his fans. (You can’t target an individual account’s fans, only those similar to them.) I also specifically mentioned emails in the call to action of the tweet, “Get soccer deal emails.”
What targeting did I use for the cards?
I’m old enough to know that you need clear targeting. For each tweet, I targeted desktop users only, male, in the USA. Then each tweet targeted Tweeters like a particular user’s followers, Robin van Persie in the one case and the US Men’s National Team + US Soccer for the other.
In hindsight, I made two mistakes here.
The first is that Twitter itself tells you that their keyword targeting is better for purchase intent. So I didn’t notice that and went with users- mistake number one. Mistake number two was that I targeted the whole USA.
I think it’s fair to say that some locations are more likely to play a lot – and probably buy cleats accordingly – relative to other locations. E.g. If you’re shoveling snow 4 months a year (hi NY), you probably play less soccer. Ditto if you’re in an urban jungle like Manhattan. If you’re in the south, however… Twitter’s interface made it seem like I had to target individual countries and since my affiliate offers were only good for the US, I targeted the US exclusively. I might try looking at Google Trends data for geographies to target next time…
Let’s talk about bidding on Sponsored Tweets, while we’re at it.
I bid max $1.01 per “engagement” (hoping naively that they would treat me nicely and sell me leads for about this price) for both tweets. Twitter was encouraging me to bid more, something like $1.30 – $1.80 if memory serves, or even higher. The suggested bids are out of whack with what you need to bid to get your ad shown. They say that you’ll reach a greater percentage of your potential audience that way, but I don’t know if that makes any practical difference in ad serving.
For example, you can imagine that someone who ignored one ad shown towards the top of their feed is likely to ignore subsequent ones lower down. Does bid affect position in Twitter timelines? I guess that’s what meant by them saying you’ll be shown to a broader audience, but I don’t think their documentation specifies exactly what happens with the bidding.
For a subsequent click focused campaign, I bid max $1.85, about mid range when they were suggesting $1.50 – $2.20. Unsurprisingly, the average cost per engagement was higher, by 18 cents, from $0.51 to $0.69. Not that I actually care about “engagements,” but rather clicks. Here again my analytics showed a much lower number of clicks than reported by Twitter. It’s pretty odd because I don’t see why it would be so compelling to click my username… But I digress. More on that in a future post.
Twitter’s documentation for sponsored tweets and direct response
If you look at the case studies for advertisers looking to make sales, many mention non-sales metrics. They read like “engagement this,” “followers that.” There’s a place for that – at the top of the funnel. If you want to target the bottom of the funnel, these case studes are pretty telling about the difficulty of executing direct response advertising when it’s bundled with brand advertising. I found that wasted my time.
Here are a few of the more relevant case studies for direct response: Bonobos pants, Financial Times, HowAboutWe, DoubleDutch B2b app. Less useful were Asda, Dominos (come on! you guys are in the most direct response category ever!!), Cirque du Soleil, CustomMade, Etsy and others.
They also don’t talk about these advertisers’ bidding at all.
Conclusion about Twitter Sponsored Tweets & Lead Generation Cards
In total, the above test just cost $25, so I’ll be the first to admit that you can’t draw definitive conclusions from this as to whether the lead generation cards are worth it. I can imagine optimizing the campaigns to get a cost per lead (CPL) below $1, and hopefully below $0.50… but the question is whether it will ultimately be profitable when I compare acquisition and emailing costs to revenue. Currently paying $3 – $13 per email is not profitable.
For what it’s worth, when I started advertising on Reddit, my CPL was also about $10/lead, and I got it to $0.66. (They since raised their prices dramatically.)
Some other observations are that:
- these tests took a LOOOT of time to set up. Hours and hours. It’s not just writing the tweets or lead gen card copy – that’s the easy part, obviously. There was a lot of backend work to do because if browsers don’t support lead gen card functionality, you need a backup URL that Twitter can send people to. That means creating a landing page, setting up email acquisition tracking etc. In addition, being an affiliate, I had to find appropriate offers to promote and so forth. At least Twitter doesn’t lie to you like Google does and claim that it takes just 5 minutes to start.
- I dislike strongly bidding for engagement, and think Twitter would have a far more compelling product if I could bid on clicks to my website or leads directly. Twitter can still guard users and the user experience from poor quality ads, simply by using better tools such as ad pricing and distribution caps. E.g. If you’re doing badly with users, your cost per impression goes way up, and if you do super badly, your ads show less and less.
- There are a lot of different targeting settings to experiment with. Besides for the different geographies (states, cities) I can target, there are also possibilities with mobile, targeting women etc. Part of this is that I’m trying to target people similar to myself, and cut corners on research.
If I want to do this seriously, I need to spend a few hundred dollars, measure and optimize carefully. If you’re interested in reading such content, comment below and/or submit this post to Inbound or Growth Hackers, and I’ll write a followup.