It's been a dismal week for social media companies that aren't Facebook.
Last night the professional networking site LinkedIn posted a disappointing revenue forecast, sending its shares plunging 25% in after-hours trading. Earlier in the week, Twitter had also shocked the market by missing revenue forecasts.
Both companies have sky-high valuations in what is a pretty frothy technology market, with investors betting that they will grow very rapidly and deliver huge profits in a few years' time. Now, both Twitter and LinkedIn are still growing users and revenues - just not at the rate that the markets demand.
The other difficult question for Twitter in particular is whether it can ever convince enough advertisers that it is the place, rather than Facebook, to reach an online audience. When Facebook floated a couple of years ago, it warned that its ability to make money from mobile advertising was uncertain - and back then you might have thought Twitter was better positioned for the mobile era. But today mobile revenue accounts for 75% of Facebook's very substantial earnings.
Twitter's problem is twofold - it has a much smaller audience, and is far behind Mark Zuckerberg's company in developing its advertising proposition. For years, it seemed determined not even to talk about a business model, now it's desperately trying to convince companies that it offers all sorts of attractive ways to reach consumers.
But how do the two networks compare in terms of the value they offer an advertiser? We decided to try the world's least scientific experiment for today's edition of our weekly radio programme Tech Tent. We bought adverts on both Twitter and Facebook to promote the programme. (I can already hear complaints about a waste of money, but it's come out of my own pocket).
On Twitter we bought a promoted tweet, to run for 36 hours, designed to increase the followers to our @BBCTechTent Twitter account. Setting up the advert, we were encouraged to target it by location, but discovered there was a fairly limited range of options. We could target Indiana, for instance, but not India or Nigeria. In the end we aimed our tweet at the UK, the US, and South Africa.
When we came to buy an advert, designed to get people to like our Tech Tent Facebook page, far more targeting options were on offer in terms of location and interests. We ended up sending our advert to 18 to 65-year-olds interested in technology in the UK, US, India, Kenya, South Korea and Nigeria.
The results to date in what I repeat is deeply unscientific research seem to show that Facebook is more effective. So far, we have spent $9.38 to generate 296 "likes" for our page with an advert served to 2,580 people. By contrast, we spent £20 on our promoted tweet.
Twitter's ads analytics report tells me the tweet received 56,295 impressions and generated 81 "engagements" - which as far as I understand means anything from a retweet to a follow.
In any case, our Twitter followers have risen by just 643 since the ad went live, which doesn't seem a great result.
Now, it's worth remembering that Facebook "likes" are a notoriously dodgy currency. As I found when I ran my Virtual Bagel experiment a couple of years ago, you can easily get thousands of likes from people who are suspiciously promiscuous in their interests.
But the same may go for Twitter followers. Advertisers are looking at the ability Facebook offers to target either a global audience or a group of 18-24 year olds interested in motorcycling in Ruislip. And, for the moment, they are betting that it offers better value than Twitter.
Tech Tentis broadcast on BBC World Service on 1 May at 16:05 GMT - or listen on BBC iPlayer Radio