I love Twitter. But their stock has been hammered lately. It has lost half of its value in the past four months alone. Meanwhile, Facebook is down a bit, but doing ok. Stock value is (theoretically) based on future revenue. If investors think a company is going to make more money in the future, they buy that company’s stock, and the value goes up. If they think the company is going to make less money than expected, they sell the stock and the value goes down. Sometimes investors get it wrong, sometimes they get it right. In this case are investors getting Twitter right, or wrong? It’s hard to say a company that is doing $2.5B and had 61% growth between the first and second quarters of 2015 has a revenue problem, but as the Wall Street Journal put it:
The San Francisco social-media company said second-quarter revenue rose 61% to $502.4 million compared with $312.2 million a year earlier. That far outpaced Twitter’s projection and analysts’ expectations, though it marked Twitter’s weakest quarterly revenue growth since its 2013 initial public offering.
Based on my recent experience using Twitter ads which I detail below, I’m inclined to believe, though not convinced, that investors are right to devalue Twitter’s stock.
Background
I run a digital marketing agency. One of our clients is Beardbrand, which sells grooming products for men with ample facial hair. We recently proposed a marketing campaign to them, the results of which I have detailed in the post I Posted Something On BuzzFeed. You Won’t Believe What Happened Next. We created images of 10 U.S. Presidential candidates with beards, posted it to BuzzFeed’s community section, and then used social media amplification to get the post moving. Because the campaign wasn’t endorsed by Beardbrand, who eschews anything related to politics but said we could post it as a “rejected marketing campaign idea,” it ended up being a non-marketing, marketing campaign that turned into more of a personal experiment where I was free to test out various marketing tactics to see what would work best.
The Twitter vs. Facebook Experiment
I noticed early on with the Beardbrand experiment that Facebook was outperforming Twitter by a wide margin in terms of generating traffic to the BuzzFeed post. But that might have been because I was buying ads on Facebook and not buying ads on Twitter. I decided to let the ads stop on Facebook for a little while, and then spend $100 on Facebook, and $100 on Twitter, and see what kind of return I got from each service. I actually ended up trying out two types of ads on Facebook, so I spent $100 on Facebook, then $100 on Twitter, and then another $100 on Facebook. Ok, I actually ended up accidentally spending $200 the second time on Facebook, but as you’ll see, the results are still clear. Here’s what happened.
That graphic pretty much says it all. I got large spikes in traffic from Facebook, and no noticeable spike from Twitter, given that the traffic being shown where the Twitter circle is located is what normal traffic has looked like on a daily basis.
What is Twitter’s revenue problem? I get more results from spending a dollar on Facebook than I get from spending a dollar on Twitter. Where am I going to spend my money on future marketing campaigns? Not on Twitter.
The Details
Here’s what the Twitter campaign looked like:
I wanted to buy clicks to the BuzzFeed article. It cost me $0.96 per click on Twitter.
Here’s how the Facebook campaigns compared:
Clicks from the first Facebook campaign cost me exactly 1/3 what a click from Twitter cost me. For the other campaign, I was buying post engagements, which doesn’t necessarily mean a website click, although 905 of the 1,023 interactions were website clicks. In other words, each click from the post engagement campaign cost me just 11 cents, or nearly 1/9 the cost of each click on Twitter.
But it would be premature for me to measure success based on what I paid per website click from the campaign, because that only shows part of the picture. In the case of Twitter, the 104 clicks are all the clicks I got from the entire campaign. But Facebook’s reporting only shows me the clicks for which I was billed. If someone were to reshare my post, and then it gets clicked on, that wouldn’t show up in ad reporting. And that’s where the real difference exists between Twitter and Facebook.
Unfortunately BuzzFeed doesn’t give me the granularity in reporting I’d like to get from their system, but the results are so stark we don’t need it. This graphic shows performance of the BuzzFeed post during its lifetime, including referral traffic sources:
Yes, I’ve spent more on Facebook than Twitter during this campaign. I’ve spent a total of $753.36 on Facebook during this campaign. I’ve spent $100 on Twitter. I haven’t spent even 10x as much on Facebook, let alone 76x as much, but I’ve gotten so much more value from Facebook there’s no comparison. Dividing the spend by total traffic from each source, including viral “free” traffic, I’ve spent 0.01 cents per click on Facebook, vs. 96 cents per click on Twitter, making Twitter advertising 9,600 times as expensive for me on this campaign as Facebook.
But, but, but…!
Yes, I know. Maybe I didn’t set up the Twitter ad the optimal way. Maybe I could have made the ad better for getting viral sharing. I admit–I’m not an expert when it comes to managing social media ad campaigns. There are people at my firm who are, but this was just me having fun with an experiment. But that’s the point–a lot of people using Facebook and Twitter for advertising also aren’t experts. If their first experiences with Facebook and Twitter are similar to mine, which network are they going to use in the future? Twitter’s challenge from its inception has been that it doesn’t immediately make sense to normal people. If their ad system suffers from the same problem, that’s a recipe for limited revenue.
More Disclaimers
This was not a very scientific experiment. In addition to the problem noted above, here are a few more problems with it:
- Facebook and Twitter have different audiences, and just because the single piece of content I posted did better on Facebook, that doesn’t mean a different piece of content wouldn’t do better on Twitter.
- My sample size wasn’t very large for this single article. If I were to spend $100,000 on each platform over a longer period of time, instead of just a few hundred dollars, the results would be more trustworthy. As it is, the post is subject to manipulation. Case in point, even though no ads have been running for the past few days, there was a huge spike in Facebook traffic over the weekend. I have no idea, yet, where that came from.
- I only experimented with a single article. Doing the same experiment with 10,000 articles would make it more than just a largely anecdotal case study.
I’m probably missing a few more. Or as Twitter investor Chris Sacca put it after reading this post:
@joshsteimle There are so many factors that would influence your example that it’s not really worth diving into. No hard feelings.
— Chris Sacca (@sacca) August 10, 2015
Fair enough. Still, based on the huge disparity between results, I would be inclined to use Facebook as my go-to for the next social media ad campaign I run, rather than Twitter. Are other advertisers having the same experience? Would most amateur advertisers do the same? And therein lies Twitter’s revenue problem, or opportunity, if you prefer the more positive spin. How to fix it? That would appear to be a problem for Jack Dorsey to figure out. Suffice it to say, Twitter needs to figure out how to provide more value to advertisers, and that means making it drop-dead easy for advertisers to make their tweets go viral, as Facebook has done. If that’s just not possible due to inherent limitations of the system, then Twitter has an inherent challenge when it comes to making money, at least relative to Facebook, whose 2015 Q2 earnings exceeded $4B. Perhaps there’s nothing wrong with that–they can still be a multi-billion dollar business, but they’ll always be a minor player relative to a giant like Facebook which can provide more value to its advertisers, and therefore derive more revenue from them. A standalone app, like Instagram, rather than a platform like Facebook. But again, we’d all be happy to own a chunk of either Instagram or Twitter.
I would love to see Twitter solve this challenge. In part because as an advertiser I want a platform I can get more value out of, and because as a Twitter user I love the service and want to see it thrive.
What do you think Twitter could do to improve their advertising system?
Liked it? Share it!
Interesting experiment, Josh. As Chris said, there are many factors that make campaigns a success or a flop.
Do you have any comparative data on user behaviour once they land on the site? Time on page, number of page views etc. That might add a greater depth to your experiment and some more insight in to the value of each – I’ve been surprised by some of that data before!
Since they’re not identical targeting criteria, it’s difficult to get a scientific, fair comparison of one platform versus the other, but maybe that might make for interesting debate.
In fairness, we’ve achieved successful outcomes from both platforms. We’d usually be picking the platform and targeting opportunities which best suited the objectives of the project – sometimes that’s Facebook Ads, sometimes it’s Twitter Ads. Sometimes it’s neither 😉
I haven’t examined all the data so I’ll admit that from many perspectives it’s a weak case I’m making. The only reason I think it’s worth asking the question is because of the vast disparity in results. I’d also love to see if others have done better research with substantive numbers. It would be interesting to know whether Facebook is truly more viral than Twitter and in what circumstances. I started looking for data but go distracted by this http://www.inquisitr.com/2318666/straight-outta-photos-go-viral-straight-outta-somewhere-website-helps-twitter-instagram-and-facebook-users-make-funny-photos/.
Ah, distraction. The greatest gift of the Internet 😉
I think it’s only really a fair test for each project in isolation. Sometimes Twitter suits and works, sometimes it’ll be Facebook. Sometimes, particularly with “viral”, it can just be coordinated timing and luck.