As you build your business, you’re often looking to improve the percentage of possible sales you close. In social or networked services, you can also take advantage of the opportunity to build repeat visits — repeat interactions with your service. This lets you:

  • have more opportunities to close the sale, if you didn’t manage to the first time
  • build repeat sales
  • create habits where your customers continue coming back to interact
  • learn about other customer needs, and send that feedback into the product pipeline
  • build a customer lifetime valuation model and use that to grow your business’s valuation

As you try to increase repeat visits, you’ll want to track and use A/B testing to measure if an idea really produces results, or just adds more technical debt for your team to maintain.

Here are a few tactics you can play with:

  • Create a weekly email — you can personalize it with member stats, and personalize the subject line, not just the body
  • Notification emails — social interactions with other members can give you a reason to jolt them with an email that can bring them back
  • Give your members something once per day… This could be a social thing, a free contact to another user, a free badge, a badge of recognition if they visit once a day for a whole week/month
  • Have statistics people will want to check regularly, e.g. # of visits to a member’s profile page
  • Think of focusing on building weekly usage first, and then focusing more on daily usage
  • Have something for your whales — your 20% of members who contribute 80% of your revenues.
  • Not all repeat usage is equal: repeat usage for potential
    repeat customers or recruiters is most desirable. As you build out prediction/valuation model for these, you’ll
    be able to test, target and measure tactics for identifying, developing and bringing this group back more.

It really pays off to test these changes. I just recently ran a test on Ryze where I expected that a daily list of new business articles could build repeat usage, but it actually lowered it. Soon, I’ll be testing out a new weekly email that’ll go out to members who’ve recently used the site. Of course, that’ll be A/B tested, so only half of the candidate members will receive the email and we’ll track which group comes back to the site more.

Keep in mind that you don’t need 2 new versions to test out — you just need one new version to test against the current version. You want to test against the current version so you can make sure you don’t make things worse. If you do want to test 2 new versions, you can test both of them against the current version, of course :).

I hope these quick notes give you ideas. I look forward to your comments and questions.

Adrian Scott is a pioneer of social networking, and is currently helping several high-tech/media companies build their businesses through metrics and testing. In the past he has also built out technology products for startups and corporations, developing software and setting up operations. He can reached via email at adrian at adrianscott . com.