Taylor Pratt, Search Marketing Specialist, nFusion Group
William Leake, CEO, Apogee Search
Brian Massey, Conversion Scientist, ConversionScientist.com
First up is Taylor Pratt
Starting with a classic quote from Avinash: “unless you’re conversion rate is 100%, you’re never done testing”
Know your users. Understand the needs of their visitors
- What motivates them?
- What are they there to accomplish?
Take into consideration: Types of personas/personality types
Testing Calls to action:
Eliminate guessing. Test the CTAs: location, button shapes and colours, texts vs graphics, button and link copywriting, CTA size
Consistency is important for a site that has a lot of repeat visitors.
Testing page layout
- The five second test a new layout: location of elements above the fold, one two three column, image location, more/less headlines
- Font type
- Making text and images clickable (use crazyegg)
- Colours: background and font
Use crazyegg a lot – find out where people are clicking, then figure out what they want to see, and make it so.
Testing site navigation:
- Is it obvious where you should click?
- Is your navigation descriptive, do they know where they’re going
- Adding active verbs to links
- The order of navigation
- converting headers to links
- breadcrumbs on or off?
- top or side navigation?
- Testing product presentation
- Ensure product pricing is clear
- adding product reviews
- adding a video: instead of or to compliment an image
- adding a call to action to each product page
- adding multiple views to a product:larger pictures, product in use
The product in motion can be great for conversions. Amazon lets you ‘flip through’ the book. They are constantly tweaking their conversion process.
Testing the purchase process
- Adding a progress bar to your checkout process!
- offering free shipping
- reducing the number of steps and form fields
- adding approximate time of arrival
- removing promo code box (you mean there’s a promo I can get!??! I need that discount, so I’m not buying now!)
- removing requirement to register
Best resource available: your users. Do user testing – simply watch people interact with the site. Make them speak what they’re testing.
@pratt on twitter
Next Speaker: Brian Massey
Okay I lurve brian, we’re playing WHAT TO TEST? THE CONVERSION GAME SHOW?!? Simply awesome.
Dang, I thought the whole presentation would be set to game-show music.
Ground rules: your business has only 3 or 4 customers living at a thousand different addresses. There are ‘profiles’. Whatever is important to your business, you can model these personas. Thumbnail personas people.
Compare Boomer/silent, Gen X, Millennial.
Tv at the left end of spectrum (boomer), email middle, social media on the right side (millennial audience).
Myers Briggs personality types apply here: Competitive, metohodical, spontaneous, humanist. These people are either emotional or logical. They make either quick decisions or slow decisions.
Scenario: business website is a plumber. We’re targeting a GenX female, 30-40, new wood floors, sink leaking, emergency
What to test?
• A link to a contact us page
• A phone number
• An EMERGENCY hotline
• A branded phone number (800-stop-leak)
Test 3 and 4
Scenario: sports equipment. Persona, male genx, spares no expense
Different navigation items to test:
• Soccer equipment
• Hot new products
• Editors choice
• Best deals
Test 2 and 3, maybe even 1
Business: manufacturer: persona, boomer, operations manager, methodical, non technical looking for workflow software
Test sales copy:
- on demand software provides best-of-breed blah blah blah
- new IBM case study, business transformation
- a long-assed page
Test 3, also test 2
William Leake is now taking the floor
Clicks and traffic? No way, we’re sophisticated and are focused on web forms, right?
dang straight. If that’s what I’m paid on. I don’t think that’s where he’s going with this…
Primary conversion approaches:
qualitative: heuristic (rules of thumb)
- top ten lists
- best practice
- a/b and analytics
- multivariate: google, omniture test and target etc.
There are many roads to conversion testing. How to market to French people? The heuristic approach would go and find a best practices list, finding things like, French people like pastel colours. The quantitative approach is ignorant of why the French like pastels, but they included the pastel page in the multivariate test, so now you know.
what does your CEO really want to get out of marketing?
- top line revenue
- bottom line earnings
most likely, not web form fill-outs. Great web-forms don’t always create good leads. A web conversion is an intermediate event, in a longer web-conversion process.
All leads and sales are not created equal. It all goes back to the funnel.
- B2C: move beyond sale to lifetime value customer (cCRM)
- B2B: ensure that you are integrated with the sales automation / crm system
- Both: change your optimization focus to “post click” and “post web”
Do you care about all personas? how about coupon clickers? The answer might be no – these can be expensive, non-loyal people that aren’t worth your follow up effort.
Alternative conversion types: conversion doesn’t have to occur on your website
- track phone calls from search
- integrate chat to drive conversion
- add coupons to in-store items
Q: if we’re starting a new site, and have no data, how do I create personas?
A: Personas are a bridge between research and people integrating things. You still need to do research, or why are you starting a site/business?
Q: if we want to not offer a promo code, but we need one, what do we do?
A: Jetblue offers an incentive at the start of a funnel, and only include the promo code box when someone starts the funnel from the incentive
Q: Tracking numbers – we like to use them, different for each campaign, but there’s a lot of leakage as people surf around the site. We had to pull a phone number from every page but the landing page. What do you think of that?
A: Try Google’s tool to generate 800 numbers just for your website. Or use third party tools like clickpath, and a few others that are phone-analytics links.
Q: if stuck to pick between a qualitative or quantitative approach?
A: The quantitative approach requires traffic, so if you don’t have a lot of traffic use qualitative. If time is the issue, use qualitative heuristic persona based stuff because you don’t have to start from scratch, and it will put you further ahead. Multi-variate has historically been expensive, not so much these days. But really, once you’ve figured out your personas, you have more data to test with, and better creatives to begin with. Generally a fan of starting with heuristic if you’re budget limited.