Rapid prototyping: Fake ads, real insight

Improving your best bet with Google adwords

I wanted to share a quick way of building up a picture of user interest and engagement really early in the product design process, pretty much at the an idea popped into my head-stage.

In a nutshell: if you have an idea for a business, service or product, you can test audience interest by taking out a small number of differently worded Google adword ads, bidding on relevant keywords, and seeing which one performs best. If this is really just an early idea, then it’s not so important where the ads lead at this point. You are testing “what if this existed”, in a lo-fi way, even before you start sketching and you’d be getting feedback within a few hours.

It’s not new, but it’s a useful method to reduce the amount of uncertainty in the product hypothesis or to increase confidence in your “best bet”. You can even do this without any branding (it changes the outcome somewhat, depending on how trusted your brand is). If this is at a really early stage, and I believe that’s when it’s most effective, then you are essentially running a fake adwords campaign. The product doesn’t exist yet, after all. In order not to waste your users’ time completely, it might be worth putting up a sign up page to let them know if and when you launch the product.

I won’t go into metrics too much at this point, there is plenty of information about that out there, but please keep in mind that the campaign needs to be geographically targeted, otherwise it may be difficult to extrapolate your conclusions about the level of interest.

Better chances of making good choices

Visualisations help, so if you’ve ever used the cone of uncertainty to show the amount of unknowns, then you’ve just narrowed the uncertainty down considerably. If you are working in a non-agile business, then this is also a good way to move from a stage that’s essentially driven by competing opinions, to the more mature stage of testing assumptions and using metrics to inform next steps. It’s of course also an opportunity for you to test your own assumptions.

Based on your learnings from this test, you can either drop the idea, rephrase or focus your hypothesis before going any further.

You can use this method as part of comparative AB-testing through google adwords to try out different versions of describing your product or service, but I’d recommend that as a second step, once you have reasonable grounds to feel confident in this best bet of a direction for your product.

User centred designers and product people have added this to their toolbox, as a way to test ideas early, cheaply and quickly.

Contexts change, the principle of testing ideas does not

This is a useful method not just for product development. Marketers use similar methods extensively and the author Timothy Ferriss used it to decide on a title for his book.

I’ve worked for a number of non-profits where social change campaigners have used this to find out which aspect of their campaign or policy programme would appeal to supporters and audiences. Service designers are using this method to test whether users would search for and accept if an element of their service’s user journey was made digital and automated (and any other number of things).

Testing seriously early is not a luxury

Working in an environment where resources are restricted and/or where the environment is changing, launching a new product or service into a noisy environment, where millions of words and pictures are competing for users’ attention is difficult and risky.

If you are the steward of an idea, then you owe it to your users and the business to test and prod the idea as early and best you can to reduce that risk.

Further reading:

The first time I heard about using this method to reduce uncertainty was on the Lean Startup site.

J Rasmussen on the cone of uncertainty.

Jeff Patton on best bets in this article on dual track product development and at his talk on losing your fear of failure

A further iteration is to do design your adwords campaigns specifically for your personas, as described in Alex Cowan’s video on coursera.