Multivariate Split Testing

Yesterday, I posted up some images of Buy buttons and add to basket buttons, I feel that I need to give some more information.

The E-commerce website is selling INKJET CARTRIDGES, Ok this is important, because ink is a stress purchase, you don’t sit at your computer and think hmmm I’m going to treat my Inkjet printer to some new ink. No you usually have just printed out a document and half way through the quality slowly disappears and you think I need INK NOW !!

so with that in mind here are some of the things we have tested..

Straplines :

Buy one get one free : yes, people still want a bargain
Next day UK delivery : this is important… it’s a stress purchase
Great prices delivered Next Day : a happy mix of both

Buttons :

This must be a UK thing, in the UK, BUY works best at the moment, I think it’s because in really life if I add something to a shopping basket ( btw they are trolley’s in the UK, not Carts ) I have already committed to the BUY. But I i’m still running the test, so we will see.

Data :

The follow up email, when spilt testing send a follow up email asking why they purchased from You. This can really help with straplines, it gives you a really good idea why people purchased, Things like :

“you had it in stock”
“great price and next day delivery”
“I don’t know hahaha, I guess I felt secure”
“I saw your brand at a friends house”
“Friend told me to use you”
“Easy to find, Easy to buy”

Being too Close to your industry :

Get your wife, girlfriend, mother, father, friends anyone not in your industry to look at the site and ask simple questions like.

What does this site sell or do ? ( don’t help them, don’t even point out which button to press ! )
How do you get to buy this product ?
If you had to use a search engine to find my product what would you search for ?

and just listen to the answers, no prompting or helping them, in the real world you can’t be there with every cusmtomer.

DaveN