Focus social bookmarking on Email, Facebook, MySpace, and Digg

Social bookmarking widgets have appeared on many blog posts and other web pages. It’s understandable why: they drive traffic to your site, and they increase inbound links and thus PageRank.

But which social media sites should your social bookmarking widget include?

After all, there are dozens of options for sharing. In fact, Sociable supports over 100 social media sites.

It’s a dilemma for online marketers: present too few choices, and you risk missing a popular bookmarking service. Provide too many, and you will confuse your visitors.

ShareThis just shed a bunch of light on this question with this report. The highlight is this breakdown of social media services by popularity:

ShareThis breakdown of social bookmarking popularity

The most popular social media services are:

1. Good ol’ email, with a whopping 57% of usage.
2. Facebook, with 21% share — and the fastest growing service
3. MySpace, with 5% share but declining
4. Digg, with 2% share
5. StumbleUpon, Twitter and Technorati all have <1% share

(Update: turns out that AIM was not in the top five, based on corrected information from ShareThis. So I’ve removed them from the list above.)

This data is surprising. First, where’s LinkedIn? (Perhaps people don’t want to share interesting articles on a site used primarily for professional networking.) Why is Twitter so low in the rankings? (For all Twitter’s press, it’s important to note that its traffic is still well below that of Facebook.) And MySpace is hanging in there quite well.

So, when you set up your social bookmarking widget, focus on the most popular services: Email, Facebook, MySpace, (maybe) AIM, Digg, StumbleUpon, Twitter, and Technorati. Of course, this breakdown won’t work for all types of visitors. For instance, technical audiences would probably want to post to Slashdot and Reddit.

What do you think? Is the ShareThis data valid in your opinion? Do you track which bookmarking services your visitors use?

What early adopters really do at their computers

TechCrunch has a great post on how early adopters spend their time on their computers. What’s fascinating about this is the time usage stats are based on actual behavioral data, so it’s very accurate, relatively speaking.

No doubt people are drawing all kinds of conclusions from this. I thought I’d share mine, which revolve around how this data affects market share, revenue, and valuations:

• Gmail is used 3x more than Google.com. Not surprising when one considers their own workday activities. But, assuming clickthroughs are more or less equal for both (valid assumption?) — that means Gmail generates the bulk of Google’s Adwords revenue. Pretty amazing considering that Gmail originally came out of a developers "20% time" project. This supports the notion that sometimes the best projects come out of skunkworks.
• Facebook is accessed 50% more than Google.com. Maybe that $15 Bn valuation is justified after all!
• Outlook is, unsurprisingly, the most used app. Now, think about Xobni. If you got it installed, whenever you use Outlook, you’re using Xobni. That means Xobni could soon become one of the most widely used apps around. That presents some interesting monetization opportunities when you have that many user attention minutes. It will be interesting to see what the future holds for those guys.
• It’s surprising that Yahoo Messenger has such low usage. Last time I’d looked, a few years ago, Yahoo had many instant messenging users. One more thing for Jerry Yang to worry about. I’m also surprised that Meebo is at the bottom of the list.
• OpenOffice and Google Apps have very low usage. For instance, Google Docs has 3% the usage of Word. If even the early adopters aren’t using them, I guess it will be some time before they start to challenge Microsoft in terms of market share. And it will be some time before Microsoft profitability, largely driven by the Office suite, starts to suffer and drag down Microsoft’s valuation.

A note on accuracy: I’m sure some will quibble about the accuracy of the numbers given that
the sample was self-selected, But market research is an inexact science.
Not to go all Rumsfeld, but you need to use the data you have, not the
data you wish you had. This, as far as I know, is the best data we have on what people actually do on their computers. (If you know of a better data source, please let me know.)

A better way to gauge software release readiness

Online surveys are, in theory, a great way to gauge software release readiness. One would think it would be easy to send a survey link to all of one’s beta customers, asking them to rate the overall stability and individual new features in a release.

Unfortunately, it’s not so easy.

Not because online surveys are hard to create or expensive. Sites like Surveymonkey make it easy to create online surveys for a very reasonable fee.

The problem is getting users to actually fill out the surveys.

My own experience is that only about five percent of beta users fill out online surveys.

Why does this matter? Practically speaking, statistical significance kicks in around 27 responses. (That’s the rule of thumb taught in market research classes. Feel free to dig into the math if you want.) Dividing 27 by 5 percent means you need 540 beta users if you’re going to get a reasonable amount of certainty around release readiness.

Getting a few hundred beta users is not easy, given the fact that beta timeframes are often crunched down to the bare minimum time, caught between engineering’s inevitable release slips and sale’s understandable desire to start selling the new product as soon as possible. (I’m not complaining about engineering here — software development is a hard activity to do, and even harder one to forecast.)

Even if you have 500 or more beta users, it make take a couple of weeks to get to 27 responses. 15 might come the first week, 10 the next, and so on.

This slow accumulation of responses makes life hard for a software product manager. Every day that goes by is lost revenue, but you don’t want to pull the trigger and go GA without having met your release criteria.

How does a product manager address this?

The easiest way is to improve survey response rates. If you can get response rates to 10%, you only 270 beta users to get 27 responses. If you can get response rates to 30%, you only need 90 beta users.

One way to do this is with a contest — for instance, raffle the hot gadget of the day (iPod, iPhone), or give away a free license of your product.

Another, cheaper way is to use "inline surveys" that appear  right on your company’s home page or blog or some other highly-visited web page.

If beta users see a blog entry with a short, five question survey, they’re fairly likely to complete that survey because they know it won’t take up much of their day. If they see a Surveymonkey link, they have no idea how many questions are involved, and they bail.

This is why I’m excited about a new inline survey from a company called Vizu. They let you embed a polling widget into a web page with a bit of JavaScript. Their widget is technically the same as other kinds of web widgets that people put on their blogs and Facebook pages, and fairly easy to setup. Pricing starts at free.

[Disclosure: I'm friends with Dan Beltramo, Vizu's founder.]