What I learned in a presentation by Greg KELTON at Orange Concept Store, 2009.06.30?

On 2009.06.30 I took part at a presentation at Orange Concept Store by Greg KELTON. I was given a prize for an online contest, and at the event there were lots of presentation. I’ll present in this blog post what I took home from his presenatation.

First of all, here are some pictures with the participants:

Participants included: Ionut OPREA, Radu IONESCU, Razvan EXARHU, Robert ANGHEL, Victor KAPRA, Radu GEORGESCU, Auras MIHAI, Constantin COCIOABA, Mihai IRINEL.

Here are some pictures with slides from the presentation:

Here is what the slides mean:
- The very first two slides came from a previous presentation, but I thought they contain useful data, and I put them in here also;
- The presentation had the following parts:
a. Quiz;
b. Why Multivariate Conversion Analysis;
c. Case Studies (Orange, Delta, etc.);
d. How it works;
e. Q & A;
- There is a slide with a comparison of two web sites; One is more cluttered, has two columns and three rows, another one has only one image;
- He asked us what do we like; Picking the example with only one image had a 75% increase in Clickthroughs;
- What would an optimal solution for picking a thing on a web site?
a. Dropdown menu;
b. Displaying all the options;
On a given test, a dropdown menu was better;
- Sidebar content is also important;
- If you have a dating web site, what’s best:
a. Two persons in one picture?
b. A single person?
One test favored the b. option;
- Who should control your marketing? That’s simple – the client, not anyone else;
- It gets more and more expensive to get a customer;
- After you get a visitor, it gets more and more expensive to convert it to a customer;
- How can you make a good decision?
a. You can speculate:
i. HiPPO – Highest Paid Person’s Opinion;
ii. Intuition;
iii. Agency experience;
iv. Argument;
b. Quasi-objective:
i. Focus groups;
ii. Usability testing;
iii. Web Surveys;
- Facebook doesn’t like the redesign;
- Simplifying a form can increase the conversion by 3.86%; When you are talking about billions in sales, that’s a good result;
- To do great testing of a web page, you can put a JavaScript Callout to a page to create a nice way to show the page to others;
- You should run experiments to your web site, even if you have low traffic;
- Repeating solutions of optimization is a great way to get results; Continue to test forever; Customer is always changing;
- Why Multi Variable Testing?
a. Listen to your Customers;
b. Improve revenue consistently;
c. Avoid cost and risk of redesigns.