Create a landing page with split testing and conversion tracking
I'd love to hear your feedback on a new landing page creator idea I'm testing:
One of the big issues I keep seeing makers experiencing (including myself) is getting a landing page up quickly to test an idea.
Now, there are a lot of landing page templates and services out there but they don't provide proper validation tools - like split testing and conversion tracking.
Split testing and conversion testing can be a nightmare to set up and it definitely doesn't help with getting things done quickly.
Therefore, I'm proposing a product where a maker can enter some copy variations, set up some CTAs and Hackr will automatically cycle the copy and test the best combinations to generate a report on the highest performing content.
All without having to install any expensive third-party tools!
What do you think? :-)
I'll play the devils advocate here. :) I think getting to statistical significance on these split tests will be a challenge since (1) you're encouraging users to run more than two variations at a time and (2) makers are likely not going to have a ton of traffic. Since the time to statistical significance is dependent on total traffic, number of variations, and conversion rate lift, this could be a real problem and real wrench in your goal of enabling users to "quickly test an idea". Maybe you've already considered this issue?
This space is hot. You should chat with @tobiaslins about Splitbee and their analytics+conversion tracking offering. Probably good mutual learnings to be had.
Have you done any AB testing yourself?
Not really. We tweaked the language of the landing page a bit and have changed the order of our images/accompanying text, but our traffic is pretty low to where we don't really think we could reach any kind of significance across tweaks.
Probably the biggest AB testing change we've made was dropping our Free Trial in favor of a Free Tier, which we did just last Friday actually (https://www.indiehackers.com/product/reflect/finally-launched-free-tier--M245YXhH15CN1NU8y8K). But still even with this change, I feel like we need to hit a certain level of traffic to get conclusive results. Is that correct thinking or not?
Interesting and thanks for the response.
Not sure dropping the free trial counts as ab test as you don't do it at the same time.
But look forward to seeing the results, we started with a free tier and don't even require sign up to use the product.
No, you're right. It's not AB testing. Our traffic is too low to do any AB tests. But I do envision us getting some different outcomes now that visitors don't need to enter a cc.
Should work out, customers these days just want to test the product out quickly and see what is best.
That's cool and totally willing to give a try @kylegawley
I can see $19 plan doesn't include split-testing and tracking. I assume people might go for an alternative provider at a more competitive price range if they only need a landing page.
While the $29 plan - interesting. What's your plan to make it better than Leadpages for example?
What do you think would be a fair price for landing page only? The pricing is fixed so I'm open to suggestions.
The plan is to keep it super focused to indie makers and build tools around their needs. One of the features I'm most excited about is integrating with ad platforms to help people generate traffic for their pages as this seems to be a huge barrier for many early-stage makers.
We're close to this with Versoly.
Analytics and conversion tracking is done we're just testing it.
For AB testing do you think a IH can get the traffic to make it statistically significant?
Yes, I don't think a lot of traffic is needed for an IH to A/B test given that they will probably be testing wildly different messaging and positioning as opposed to optimised variants of the same message.
You might be right if https://neilpatel.com/ab-testing-calculator/ is correct.
I used another calculator and results showed you needed 1000s of visitors
I believe it depends on what's being tested (I have no research to support this).
Splitting testing a different button color with 500 visits is probably insignificant, but has relevant with 10,000s of visits. Whereas, split testing 5 very different headlines with 500 visits is.
At a conversion rate of ~5% or so? I doubt it.
How would you recommend testing messaging at a very early-stage with low traffic?
Do sales first, learn what messaging your customers use and respond to, use that on the site.
Isn't that just extremely low-volume, un-scalable split testing?
Nope it's quite the opposite.
Split testing is a push model. You write different variations of the copy, and get a binary (yes/no) answer in return.
If you do sales, you're following a pull model. Your customers are basically writing your copy for you.
If you're just starting out, there isn't a more efficient way to get your messaging good enough to make sales.
I think that makes a lot of sense, great idea! would you be able to vary the layout as well as the copy/CTA?
Thank you! :-)
I was thinking I would provide a tried and test structure for creating the content (headline, ctas, benefits, callouts, features, pricing)
Then provide some different theme/template options.
I don't want to go down the road of creating a CMS and providing a lot of flexibility because introduces too many barriers to get a page up quickly.
Makes sense! To be able to measure some kind of significance, would you also have the analytics built in to determine baselines? Also would there be a/b testing tools built in to provide guidance on how long certain variations should run?
Yes, what I want to get to is using collective data across the platform to figure out what works best for most people (how many callouts, how many ctas generate the most conversions etc) then provide guidance on how to achieve the best results based on past performance.