Web Interface - Golden Path Improvement

Embrace the future of web site path analysis and conversion improvements.

Optimizing the conversion paths through our web sites is a task we Digital Marketers tackle every day. Senior management requires ever-increasing gains in site conversion rates and visit efficiency. And there is a standard model for how we’ve traditionally done optimization. Using a web analytics tool (Omniture, WebTrends, CoreMetrics, Google Analytics, etc.), we take snapshots of site usage data either daily, weekly, or monthly, and supplement those reports with ad-hoc queries to try and figure out specific problem areas on the site. Then, we follow the standard cycle:

It is slow and laborious, is based only on snapshots of site data, with unclear results.

The result= lots of guesswork/trial & error. Your team makes changes, and then you take another snapshot of data. Maybe your changes solved the problem, and maybe they didn’t. Luckily, the nature of site development allows for lots of experimentation…so you can go back and try something else. Again and again, the cycle endlessly repeats. What if you could cut your wasted efforts to almost nothing, and truly target site improvements on the areas that will have the biggest impact, based on a more complete view of your site users?

The future of web site path optimization.

Cutting-edge E-Commerce & E-Business site leaders are taking a more sophisticated approach to site optimization. Gartner recently published the following view of web site design maturity...

As you can see, the more data you collect, the more sophisticated your site design recommendations can become. Where are you in your data collection/site analytics maturity? So the question lingers: If you need to collect and analyze more data to become more mature, how do you get access to all that data, without having to hire more people and increase your already crushing workloads? Teradata has the answer.

A new view of site optimization.

Collecting a historical view of each customer, including their current and past interactions with all channels of your company, provides a much richer view of why a customer picks certain paths and clickable elements on your site than a homogenous, snapshot view (the old way).

Running your clickstream data up against a centralized Data Warehouse that holds the records of your customer interactions is the new way to predict the most highly-converting paths through your site for different segments of site visitors.

A major Teradata customer, eBay, runs its entire decision-making platform on a centralized database environment from Teradata, with massive amounts of incoming data per hour. For example, they manage more than 167 million paid keywords, including variations, languages, rank/placement, and geography. They tie all these together to quickly analyze which combinations provide the most traffic and conversions, and their digital marketers can quickly optimize their Search media spend at a minute level, saving millions. Similarly, they run thousands and tens of thousands of tactical A/B tests at any given time, constantly optimizing the pathing through their web site and gaining data to customize their web experience.


How does it work?

Identify optimized conversion & navigation paths

  • Enables optimization of site – find Golden Path, reduce Abandoned Searches
  • Enables automated experimentation and A/B testing – a controlled environment to test alternate paths for lift
  • Flexibly navigates through unknown paths
  • Efficiently “links” billions and trillions of clicks in a single pass to find matching path results
  • Millions of rows in >> thousands out
  • Can run hundreds of A/B tests/week

All of this against massive scale: For one customer = 2 trillion rows, 20K queries/day averaging 18.8 seconds

This kind of view will change the way you look at your web site optimization…and reap IMMEDIATE rewards. So how do you get started? First, we will help you identify the types of web data that you want to collect.

  • Date & Time
  • Log-in Identifiers
  • Path Actions-all the URLs
  • Activities on-site
  • Browser type
  • Cookie identifiers

For Digital Media & Interactive Leaders, this is “Big Data!”

Then, identify the events of interest. Shopping basket. Abandoned session. Successful session

  • Run path analysis around those events of interest
  • Map the different preceding events that were captured for that customer
  • Start by including everything, and then narrow down to eliminate things not of interest, which are not important
  • Get as much data as possible so analysis can expose outliers you didn’t think were valuable

Length of time you select will depend on event. If measuring Customer Loyalty you might want to look over a long period of time; if measuring speed of conversion on mobile vs. laptop you might look at a few minutes of data. A key value of Teradata Aster platform is that it can run quickly, using iteration, rather than having to do it all up-front—and you reap the benefits of receiving automated summaries & reports to target the improvement spots most likely to increase your site’s effectiveness.

But what about Hadoop?

We view Hadoop as a very effective storage platform for the kind of multi-structured data generated by web site activity. For a deeper view of how Teradata and Hadoop technology integrates, click here. Fundamentally, the best process is as follows...

  • Store data in Hadoop
  • Bring into Teradata Aster
  • Use Aster operators
  • URL parser (parsing urls onto different page views)
  • Sessionization operator
  • Run npath on data to create clickstream path and do investigative discovery

All this is included in version 5 of Aster MapReduce

What kind of Web Site Insights about Visitors can you start Seeing?

Example: Lowe’s

Questions we can help answer from web site pathing clues...

  • How do visitors find you? arrive at web page?
  • Type in a store or zip code? Have an assigned store?
  • Do they sign in? Identify as a ‘business’?
  • Search a specific item? Did I start a basket? Or view history?
  • Do they check services? Limited-Time Values?
  • Do they enroll in email?? or Call?
  • Do they follow Facebook? Twitter? Or leave to Facebook or Twitter?

Or, Shopping Cart Sample Insights

  • So I sign in, now what?
  • Have I left items in my cart? MyLowes? Projects?
  • What are my preferences? History?
  • Did I search for something?
  • How did I shop? Most Popular, review, price? By department? Do I read reviews?
  • What is in my cart? Do I pick up in store?

Imagine what having these kinds of insights, run across thousands to millions of users, can do for your site optimization efforts. It’s the kind of analysis Teradata systems are designed to run, so you can make the very best web site optimization decisions possible. To schedule a deeper discussion with one of our associates to talk about how you can put the power of Teradata to work in your web optimization environment, click the ‘Contact Us’ link on the right side of this page. We’re excited to talk with you!

Additional Resources