SES San Jose 2009 is off to a rockin’ start and this afternoon I’m sitting in a session about turning web analytics into money.
Avinash Kaushik, from Google, shared this #1 rule with the audience:
Don’t Stink! If you do, nothing else matters.
To get started, figure out which pages Google has determined are your home pages (i.e. the pages they send traffic to) and fix those pages first. The biggest opportunity lies with these pages and reducing bounce rates.
How to Rock It!
1. Focus on user behavior
How many people purchase 0 days after visit, 1 day after and so on.
Try softening the call to actions on the landing pages, knowing that very few want to buy right now. Rather, most are researching and then will go back to the most helpful website to make a purchase. It’s not all about price!
2. Optimize Match Types
Make sure your metrics match to net profit, not revenue. Revenue can be increasing while profit is decreasing.
3. Convert Opinions into Hypothesis
Don’t let opinions prohibit your company from making a profit. Take the opinions and start testing them as hypothesis to get outside your perceptions and into data.
4. Competitive Intelligence
Find it and use it.
Bryan Eisenberg of Future Now promises to help the audience learn the secret to making analytics work for them.
It starts with prioritization and all comes down to budget. How much can a company afford to invest either in testing, creative or resources.
1. To do web analytics correctly, make a to-do list regularly
- what marketing efforts or parts of the site have challenges
- what needs to be improved
- what things need to be tested
- what efforts should be reduced
- what efforts should be increased
Then, prioritize based on the answers above, specifically which will impact objectives. The biggest failure of many companies is generating a report and doing nothing with it.
2. Segment your way out of sadness
Average metrics produce average results
Marketing efforts are meant to target different segments of a target audience and measurement should fall in line accordingly.
Understand the different types of people visiting your site and then create personas based on those different segments: competitive, spontaneous, humanistic, methodical.
Persuasion of those same people is a process, not an event.
3. Always Be Testing
A/B and multivariate testing, available with Google Website Optimizer is a great place to start.
If you can, conduct user testing to identify usability concerns and get to the voice of the customer.
Next, the GodFather of analytics Jim Sterne steps to the podium.
Tip #1 – Do what Avinash said
Tip #2 – Do what Bryan said
Tip #3 – Do it again
Beyond the 3 steps above, make sure to optimize the shopping basket. Look at Amazon for an example with 1 click purchasing (sweeeet!).
Don’t make it hard for a visitor to buy from your website.
When optimizing the basket, determine the value of a visitor. Don’t forget that a conversion might be from a sequential search process. i.e. a visitor searching for camera, then digital camera, then 8 megapixel camera, then a search for a Nikon coolpix. Yes, the Nikon coolpix converts higher – great! But don’t discount the value ‘camera’ and ‘digital camera’ played as ‘assist keywords’. They led the visitor down the road to purchase.
Q&A from the audience:
Question: When someone is coming to the site, you mentioned softening the call to action from a buy-now to a send me more information. Have you seen the conversion rate drop?
Answer from Jim: What we have trained people to do is put an item in the shopping cart, input their email, abandon and then wait 24 hours to receive a 10% discount if they come back and make a purchase they intended to make anyway.
Answer from Avinash: I’m patient. If I have to wait a couple days or couple weeks to result in higher conversions that’s fine with me. If you are only worried about selling for today (sales people!!) then you are going to lose out on the visitors who are seeking to be educated prior to purchase.