B2C Customer Development, part 3 of 3: Testing with pre-launch landing pages
Cracking the signup page for a closed beta has become something of a holy grail to many budding startups. Fueled by success stories from the likes of Hipster, who gathered up tens of thousands of emails without even telling people what they do, and the simple platforms like LaunchRock or MyBetalist for constructing signup pages, it may seem like that’s all you need to build a user base.
Why A/B test the landing page
In developing a B2C product, a pre-launch (and, let’s be honest here, pre-product) landing page can efficiently validate or falsify hypotheses, check product assumptions and even build a leads list. But if you only concentrate on the latter, you might skip the opportunity to fine-tune your offering. And this is the cheapest moment to do it, not six months down the line.
The first thing you want to make sure you have is A/B testing. Choose what you want to use before you build the page. If you build the page yourself, consider Optimizely for quick and easy A/B testing of pages or page components. If you want to use a service that includes page creation and testing, something like Unbounce might be for you. If you decide to build on LaunchRock or MyBetalist, for example, you want to make sure that in addition to just putting the widget on your page, you will include other elements there.
For example, assume you are creating a niche group-buying site (a nicely passe example). Do you concentrate on cost savings, quality of the curated offers, targeted user base, exclusivity, or what? You might well want to test a few different ways to position your product. Or maybe you have a web app for collaboration. Are you solving the problem of calendaring, coordination, communication, SMS updates or syncing across devices? Maybe a combination of the above? If you are not sure what your target audience values the most, good news – now is your chance to find out.
Expressions of interest
If someone gives you their email address you have a potential lead. But it is also a positive signal, an indication of interest in your offering. Entering an email address can be a very strong signal (and thus also the one with a higher threshold), but you should try to gather other signals too. Adding a Twitter/FB/G+ button on the page is one positive signal you want to track. Linking to more content by saying “Find out more…” will also tell you if people are interested (they might also be confused - remember to make your own hypothesis).
Make a list of all signals you can and will track. Make sure you understand what they most likely mean. Then, design a landing page that describes how your product solves a user’s problem and why they have always missed it without knowing it. Then, prepare to make variations on your design and turn on your A/B testing.
Three elements of a good landing page
In addition to concentrating on describing the problem and your solution to it on your pre-launch page, and varying this description as described above, there are three more things to consider.
1. Visuals
Make the description of your solution very visual. It does not have to be an accurate screenshot of what the finished product looks like. You can visualize in sketches, stop-motion animation or flowcharts. If you have done rich prototyping (for example, interactive PDFs) as detailed in the first blog post of the series, you will benefit from this in creating your landing page. Recycle this material. If tested with it with real people, your results from these tests will be directly comparable to what you find out with your landing page.
2. Soundbites
If you tested with real people, you may have received some good feedback. Maybe someone thought your product answers their needs exactly and said something very nice about it. I hope you replied with “I will quote you on that”, because here is your chance to do that. Testimonials, as long as they are real, honest and by real people, are a strong feature on a landing page.
3. Virality and traffic optimization
Finally, of course, the wet dream of the landing page creator – virality. LaunchRock landing pages aim at virality rather artificially, and while they report good results by their users, I would expect that the approach only works for certain types of products, and that a certain signup virality fatigue may set in among the early adopters who get peppered by these offers.
There are a couple of ways to encourage virality. One, by all means test telling people that the more people they ask to pre-launch, the faster they get access. It may work. Or you could ask them to pay with a tweet to get access (for example with www.paywithatweet.com). Finally, something that works almost like magic – asking nicely. Tell your interested user that you would really appreciate their help in getting the word out about the new product, and if you do it genuinely, they might well respond.
Sources of traffic
Once you have the pages up and A/B testing running, where do you find the people? Listing in places like Hacker News and Betali.st can be worth it, so explore all of these free channels. In paid channels, Google AdWords is excellent as it allows you to test your message also outside of your own properties (which warrants a post of its own). Facebook Ads can be a cheaper way to get semi-targeted traffic, but also look at other platforms, like StumbleUpon ads, where you can get visitors from $0.05, which can also help you organic StumbleUpon placement (though be warned, they serve their traffic in spikes, so brace those servers). The beauty in this semi-targeted traffic is that your A/B testing landing page works harder for you. Remember the point about aiming slightly wide with your target audience in the previous post of this series? The same principle applies here too, and should apply as long as you are testing.
The previous posts in the short series were about prototyping and about finding users to test with for your B2C product.
Mikko Järvenpää, Marketing Geek, HackFwd
