This is a live blog post by guest blogger, Imelda Khoo. Imelda is the E-Marketing Manager at Tektronix, responsible for global SEO, PPC and social media. Imelda blogs at SEM Booty and is also on Twitter @imeldak

It's time to get down and dirty... with data.

We have Bryan Eisenberg moderating this session, and you will respect his authority. The speaker line-up is just as impressive: aimClear's Marty Weintraub, ROI Lab's Tami Dalley, and SiteLogic's founder Matthew Bailey.

If you haven't seen Marty in action before, you're in for a real treat.

Here we go.

Marty is up first. He's going to show us conversion reports your boss can actually understand.

Marty says analytics aren't fancy - It's how you organize and how your look at the data that matters. Bosses understand "ker-chiiing", so make sure you focus on conversion.

Conversion trends can get lost in the noise. He's using Advanced Segments in Google Analytics to get to the good stuff.

Try breaking out events in Google analytics by conversion to find some interesting data, including:
- Conversion by source (direct traffic vs. search vs. referrals)
- Conversion by geo
- Conversion by time of day (good info for your community manager)
- Converted YouTube user profiles (and friend these people)
- Conversion by new vs. returning visitors
- Visitor Loyalty can give insight into attribution
- Which browsers convert (and triangulate with source)

Marty's main takeaway: Don't let conversion get buried.

Matthew is up next, looking at three obstacles to deep analytics.

The three main obstacles are:

1. Dashboards
Big bosses like big numbers, but these kinds of numbers (especially "hits") are not important.

2. Velleity
The desire to change, but not to do anything about it.

3. Hamster Wheel Analytics
Putting data into a spreadsheet and spending the next few weeks justifying to your boss why numbers are higher or lower than last week/month/year. You get locked into the cycle every month, and can't look ahead. Sound familiar to anyone?

The cure
Understand your customer by slicing away the big picture data and drilling down into specific customer groups.
- Find: Intent
- Determine: Expectancy
- Observe: Reaction
- Analyze: Behavior

Find: Intent
Segment your keywords, your search terms, and the ways people found your website. The key is to start telling the story of a specific group of people, then break these out still further into smaller buckets. Once this is done, you can start to build relationships based upon the data and compare intent across the different groups.

Determine: Expectancy
Determine expectancy by finding out how visitors found your site and how long they stayed. People who stay longer often do more, and have a higher conversion rate. People will engage the site differently based on where they came from. Typically people from Twitter don't tend to stay long.

The highest context sources often lead to higher engagement (in order):
1. Blogs articles
2. YouTube
3. Forums
4. Search
5. Social News
6. Twitter

Observe: Reaction
Observe different reactions in:
- Bounce rates
- Design

Analyze: Behaviour
Look at what people do and how they behave when they land on the page
- Look at specific pages
- Compare actual behavior to strategy
- Measure reaction to entry points

Finally, take action on your findings, otherwise it's all for nothing.

Last, but not least, Tami is up. Tami says she's a professional nerd and loves deep dive projects. I like Tami already. Tami's motto is "If I can't measure it, I can't manage it!". A key lesson she has is to mix it up and not to look at a single metric in isolation.

It's important to focus on a plan up front:
- Create a hypothesis
- Define scope
- Pull data
- Follow the crumb trail
- Translate into action (the "so what?")

In order to answer the question, "How should I change my market budget allocation based on geo performance?", Tami looked at data split by region, keyword cluster, and messaging for each cluster. It took many, many hours to go through, but the deep dive was worth it. When it came to applying the insights, she generated 20% more revenue with the same amount of spend. That's pretty awesome.

Things to look out for when doing a deep dive:
- Sample size
- Major site changes
- Seasonality
- Accurate tracking
- Change in marketing spend