Coremark Analytics

I first want to send a “Thank You” to the Web Analytics Association for the free pass to the Houston, TX Online Marketing Summit.  If you are a member of the WAA and don’t read the monthly newsletter, you should.  Case in point – free passes to conferences (summits, forums, seminars, etc…) and other great discounts for just being a member.    

The conference was packed with great presentations covering a vast amount of information.  Between my co-worker and me we attempted to attend each talk within both tracks.  The Houston, TX location was setup in two tracks one focusing on “Search Marketing & Website Strategies” and the other on “Email Marketing, Analytics, & Social Media.”  Below are some highlights of those talks I sat in personally – but overall the conference was great and I learned a great deal and would recommend it to anyone in the vicinity of the remaining locations. 

Highlights from the talks I joined in on:

·         Google Website Optimizer; Dave Underwood, CEO, TopSpot:  Test minor changes yet don’t test things you already know don’t work.  Top variables to test include headline, image position, ‘call-to-action’ placement/look & feel, length of page, registration requirement for downloads, and contact form field list.  Key points from Dave included listening to your audience while testing, plan ahead and identify what you want to test up front, make sure to run the test long enough yet make sure you don’t over run the test (you are only hurting in the long run if you allow the ‘bad’ versions run longer than needed), and lastly “Just Test It.” 

·         Top 10 Email Campaigns; Joel Book, Dir. Of eStrategy, ExactTarget:  Joel’s presentation focused on permission based email and covered a great deal of information that I can’t give justice in a few short lines, but here goes.  Joel stressed that email strategies should be used to maintain customer engagement with your company while Search is used to attract and the design of a landing page is to convert.  Within a plan (step 1) one should design a communication that will give your customers a reason to Opt-In.  Test, test, test and integrate web analytics to understand the whole picture.  Top five things to test were landing page copy, A/B testing the offer, Subject lines, and A/B testing email creative.  Focus on understanding what your customers want and leverage a customer preferences center to deliver customized content to fit their needs.   Joel mentioned a few supporting tools and resources that can assist in design with email campaigns – PivotalVeracity.com, EyeTools.com, and EmailExperience.org.  I haven’t personally used them but they sound promising. 

·         Beyond Google – Vertical Search; Chris Hulse, Business.com:  With no great surprise there is no common list of available vertical searches but here are some that were mentioned during the presentation – business.com / ThomasNet.com / GlobalSpec.com / CitySearch.com / SourceTool.com / Shopzilla.com / Shopping.com / KnowledgeStorm.com / VerticalSearch.com.  G Y M is the new acronym for the top 3 search engines – “Google Yahoo MSN.”  Vertical search engines should be used to enhance, not replace, online paid placement marketing.  Your marketing plan should include “Core & Other.”  Scan the landscape of your users and understand their needs and other resources they use day to day to enhance placements. 

·         Social Media – Beyond the Buzz; Jason Breed, Vice President, Neighborhood America:  When integrating social media in to an existing marketing strategy, social media should enhance, not replace traditional online media outlets.  Start out by identifying the goals and select the right technology.  Ensure that the infrastructure can support the anticipated response times ten.  Create the right environment for your community.  Develop a community that provides value for those members.  Ensure they have a common interest and that environment is trustworthy and safe.  And most importantly, establish clear expectations up front.  Measure metrics that matter which include those that increase revenue, decrease cost, and those that help drive those two faster.  Lastly, ensure scalability and reliability of the community – Repeat.  Some social media sites mentioned were MySpace.com, Facebook.com, and Digg.com. 

·         What’s next: 21^st Century Lead Cultivation; Nate Pruitt, Regional VP, Eloqua Corp:  Nate brought to the table the idea of Lead Scoring. Lead Scoring is the process of assigning a numerical value to each incoming lead that is then used to rank them for priority processing.  Developing a lead score is pretty straight forward includes identifying interest indicators that best predict behavior.  Align these indicators to lead quality and then assign a weight (positive or negative as you need accelerators and decelerators).  Nate also discussed the idea of Lead Nurturing and the process of marketing to so called “bad” leads or dead leads. 

After reading all this, you may wonder, what does this have to do anything with statistics?  Well, if you think about it… it has everything to do with statistics.  In each and every discussion, metrics/measurement was mentioned in some form or fashion.  Whatever the strategy, campaign, or initiative, measurement of success is at the heart of each and every plan.

Until next time… safe analyzing. 

… and stop misusing data mining - is Peter Fader’s message to CIO’s. CIO Insight interview with Peter highlights the strengths and weaknesses of applied data mining in the business world and I have to agree with some of his thoughts; especially on the topic of utilizing probabilities to measure the propensity of behavior.

Measuring the probability of users actions can be strong and powerful if used properly. And can be easily done in Excel.

The trap I see so many people fall into is trying to analyzing too many variables at once and not taking the time to even look at what they are throwing into the model. If you really want to you could probably find relationships between how fast the sun rises and the stock market closing rates but does that really make any logical sense? Then why would you try to build relationships between buying behaviors and the fact that they own an Apple iPhone if you are selling shoes? So many marketers want to know every little detail about their customer – demographics, psychographics, what kind of car they drive, etc…

When you throw too much data at a problem you will have a hard time with independency and you need to take careful consideration the structure of your data otherwise your predictions can lead to false outcomes.

Some rules of thumb from my perspective:

1. Enhance your data with the VOC – take surveys online or telephonically (mailed surveys are costly and too time consuming). This is a great way to get anecdotal data you don’t see in click stream data.
2. Familiarize yourself with all the variables and truly understand what they mean – not what you think they mean.
3. Don’t use variables that you can’t reproduce easily. If it’s too hard to calculate, find, or collect from the database then you probably shouldn’t use it. It’s impractical.
4. Only include variables that make sense and add questionable variables later and determine if they degrade or enhance the predictability. In the end you may not even find a reason to test out the use of those questionable variables. *Make sure to not include variables that are variations of each other. If you include % of visits this month then don’t include the frequency of visits this month too. This can cause problems with multi-collinearity.
5. Save enough data for testing! Minimum split is 90/10 but recommend at least 80/20 split. That is at a minimum use 90% of your data is used for development and the remaining 10% is for validation of the model. You need to know how predictive your model is before you take it to the market.

Bonus Point -

6. If you want to get fancy, look at a repeated measures DOE structure for analyzing transactional data.

Until next time… safe analyzing.