Hey there.
You know how the product launch process is such an arduous one, taking a long time to put together, so many moving parts and all that?
And how you’re typically completely exhausted when it’s all finally LIVE, right?
If you’ve been through a product launch yourself, been part of one, or even watched one closely, you know this is how it goes.
However… when you’ve done the bulk of the work and you have that surge of excitement of ‘It’s LIVE!’ right in your arm… that’s when you need to think fast, and think hard!
Why? I’ll explain.
When you run a product launch, you get a sudden large influx of qualified traffic = a Conversion Rate Optimization expert’s dream situation!
You will be able to get significant test results within a short period of time while the launch goes on – not weeks or months as you normally might have to.
In fact, you can probably run several A/B tests back-to-back, even multivariate tests… WHILE it’s going!
So this means – you can improve the actual sales for your launch… WHILE it’s still going!
You get to take advantage of the convergence of high traffic, qualified leads and the opportunity to test many things in rapid succession.
It’s a shame how few marketers actually take full advantage of this great confluence of testing factors, a Great Marketing Storm if you will…
Getting great sales is one thing, but getting great sales AND learning a lot about your audience and your marketing… I think you’ll agree that’s much better!
This way, your next launch just might be… wildly more profitable… not to mention ongoing profits for the same product!
When you add it all up, you know you’d be crazy NOT to split test the heck out of your product launches. When you spend so much time, effort and hard cash to put on the great show known as a product launch… For goodness sake, make the most of it! 😉
People reacted to this story.
Show comments Hide commentsI’m thinking of launching a self-help book on my blog soon. Any suggestions?
Chetan – just go ahead and do it. 🙂 What do you need suggestions for?
With regards to your title question, my answered is Yes. Having implemented with split testing is a vital part to achieve your ideal goals and success you for your campaign. Needless to say, that we can measure which part of the campaign are getting good performance and results base from your implementations.
Its very vital factor to implement split or to multivariate your campaigns if you’re running some ads on searches, needless to say its very important factor to oversee whose campaign are effective or not and to determine which is best performing.
I would like to know more about this topic. I use 1shoppingcart and it’s not easy for me to figure out how to split test. One way I’m doing it is by creating two different product codes and seeing which on sells best, but then I sometimes wonder if it’s the sales copy that’s different on page B that accounts for the higher conversion. Am I going about it all wrong? Suzanne
This article makes a great point. The biggest difficulty to any sort of multivariate statistical testing is a lack of samples. Sure if you’re Ebay or Amazon you have no shortage of experimental subjects. But for smaller sites it could take a very long time to learn much. So it makes sense that you would want to do your testing when traffic is at its peak.
Dear W – you’re right it takes a certain amount of traffic to get any meaningful results from conversion testing. In case of low traffic, there’s basically two choices to make: drive more traffic, or create a more ambitious variant to test with, so you get clearer results in a shorter time. That can work well, too.
Suzanne – you hit on a very important point. It’s crucial to know which changes affect your results, and to know when you have meaningful results to begin with. Therefore you need a controlled environment to do A/B testing, either custom testing scripts or a purpose-built testing suite such as Visual Website Optimizer. In short, you need to choose which elements you’re testing at one time (some part of copy, pricing, titles, many others) and then initiate tests one after the other. Then you’ll know where the differing behavior comes from and you accumulate real changes in conversion. I hope this helps!
It seems like this has become almost standard with product releases of any kind. Even with the Google algorithm changes, newer sites or newer products get a bump up in the first few weeks to help increase visibility. It also helps them “organize” where your site or product should be categorized under, based on the traffic response. Looks like those first few weeks of a release are crucial, and a way to help finish molding your branding to your customers.
Lauren – sure new sites with product launches get a spike then a bump.. but almost nobody actually utilizes that to split test.
Hi Juho, I’m just about to launch eshop on my personal web about art and after I read your article I’m now thinking back to roots, to launch it with the best I have, because as the launch it will probably make highest performance, than after people get used to it. Thank you for the post, maybe I will delay it little bit.
Recently I have said by my boss that you should do A/B split testing for one of the domain we have for our business. I used some tools, but I think I can not find the right one as per my expertise. I think I need to find some tool which can be fit for my needs. On the other hand I need to clearly understand what is split testing and how it can be functional using a tool.
For “Software Testing Company”: It’s not really the tools that create the results. You can have the best tools in the world and get irrelevant results still. You need to educate someone at your company to understand conversion testing or buy it outsourced. In many cases, an outsourced service is better as you get to hit the ground running.
Martin: Often it is a good idea to just launch when you’re ‘good enough’ and then test the funnel diligently for best results. Alas, almost nobody does that. I guess it’s the perfection syndrome…
Hi,
The biggest difficulty to any sort of multivariate statistical testing is a lack of samples. Sure if you’re Ebay or Amazon and ecommerce you have no shortage of experimental subjects. It was pretty much good idea.
Hi Jessica – thanks for stopping by. What do you mean by lack of samples? If you mean lack of traffic, that you can always buy from Google if it’s not coming any other way. 🙂
I’m not necessarily advocating multivariate testing anyway, it’s good for certain kinds of situations, but not always.