Research

PPC/Landing page testing

AKA Smoke testing, ghetto testing, pitch testing... In this kind of testing, you make a fake landing page for a product/service that doesn't exist, and then point Google/Facebook/other ads at it. By measuring click-though rates on the ads and also possibly on the page, you gauge the interest people have in the idea.


Good for:

  • validating that ideas for products are interesting
  • this tests real behaviour, so can give a more accurate read than asking people about their behaviour
  • testing variants of a proposition or product/service idea against each other
  • if you use Facebook ads, you can target people by their interests so you can test your idea with your precise target market.

Bad for:

  • nuanced insight: you just get what people do, not why
  • if you're testing in a popular market sector, the ads can be expensive
  • if you're testing a very niche product or small market segment, it can take a very long time to get enough of a sample size to get statistically significant results.

How to use

  • Design and write copy for a landing page. You can write several different variations of this if you want to test different ideas, or just ensure that you're not biasing your results by having a hard-to-understand landing page. It can be good to run a quick 5-second test on the landing page design, to see if it's easy and fast for people to understand it.
  • Create the landing page. We tend to use Unbounce for this, because it has the basic analytics built in, it's fast to use and non-technical people can use it. When creating the page, it's really important to make sure you've set up conversion goals for what you want to be tracked (form submit/button click)
  • Make variations of your landing page, if you want. Instead of making a new landing page, if you make a variation on an existing one, you can use Unbounce to send a % of traffic to the URL to each of your variations. Voila, multivariate testing.
  • Consider creating a second 'more info' page, which can be reached via a link on the actual landing page. This can give you a second read on how interested people are in the product/service, by measuring click rates on this button. You could even collect email addresses on the page they reach after clicking, and contact these people for depth research. If you want to do this, make sure that you make it clear what the email addresses will be used for, for example: "Give us your email address to hear more about this product, including launch dates, and give your feedback on the product".
  • Write advert copy. It's a good idea to write a few variations on this. This helps control for the fact that a particular way of phrasing the idea might put people off, or be unclear. If you're using Facebook, remember you can put text in the images you use. In fact, this text is usually the most attention-grabbing.
  • Upload the adverts into Google/Facebook, and point them at the Squarespace URL.
  • Start the adverts running, with a reasonable daily budget. This will vary depending on the keywords or demographics you're targeting.
  • If you're testing 2+ variants against each other, monitor the statistical significance (using this Google Spreadsheet, if you like). When you've got statistical significance, stop the ads.
  • For really rough guidelines, 2%+ CTR on Facebook Ads is good.

Outputs

A fake landing page

Click through rates that give you a sense of how interested people are

Possibly some email addresses of people you can contact to ask for more detailed feedback

Software/services

Timeframe

Setting this up takes 1-2 days, or even less. It depends how much additional design time you need, and how long you want to spend refining the copy.

The time you need to leave the ads running really depends on how many searches use the keywords you're targeting (if you're using Google Ads), or how many people fit the profile you're targeting (if you're using Facebook Ads). Another thing to bear in mind is the time needed to get statistically significant results, if you're testing 2+ variations against each other. To be safe, leave 5 days, or even longer if you know it's very niche.

What you need

  • Copywriter OR someone who can write copy
  • Designer OR use existing designs in Unbounce
  • Money for Facebook/Google ads. To work out how much you need, you can use Google's keyword planner for a daily cost estimate.

References

Advisors

Katy, Kat, Ian


Notes

  • x