User experience has many applications. We’ve seen people adopt it pretty aggressively for incident management and service level management. But we’re also working with customers and third-party partners on a number of other applications.

User performance data joins test-based and device-based monitoring as the three fundamental building blocks of web performance management. And just as testing is used everywhere from capacity planning to reachability monitoring to penetration testing, so real user monitoring is finding a wide range of applications.

One of the reasons for this is its relevance to groups outside of IT. Business information such as the value of a transaction or the name of a subscriber are a part of the data that’s collected, so it’s much more than just performance information. It’s a real-time feed of user activity that gives the business insight into its online interactions.

I put together the circle diagram below to illustrate some of the ways that user experience is being employed.

Starting with the fundamentals — good, accurate, detailed per-hit and aggregate data collected from not only web pages but also Rich Internet Applications — user experience applies to all of these areas:

• User Analytics, in concert with a web analytics tool to look at conversion and search engine sources. For some web applications, user experience is the only way to collect transaction information since the site isn’t publicly deployed.
• QA and testing, both at the start of the test cycle (recording a user session for later use in a load-testing application) and at the end (watching code as it goes into production to see if QA missed any issues.)
• Helpdesk, for problem diagnosis and user assistance.
• Billing, for generating usage reports by subscriber or customer and assessing bills for excessive use.
• Dispute resolution, using facts instead of anecdotes to see what really happened and resolve an issue fairly.
• Incident management, in which problems are detected as soon as a user experiences them — before the phone rings — and resolved using the forensic data that was recorded from the web session.
• Service Level Management, generating performance and availability reports by customer, geography, or branch office.
• Baselining, watching a particular function, server, or site to get an idea of what “normal” is in order to set thresholds or measure long-term growth.
• Capacity planning, in which the relationship between traffic (load) and latency (performance) is calculated over time to see how much a site can handle before becoming unacceptably slow.
• Compliance, keeping a record of transactions for long periods of time in order to comply with industry law or regulations or to protect the company from risk.
• Fraud detection, in which user traffic is analyzed to look for patterns of anomalies or inappropriate use — from hack attempts to site harvesting to sharing of account logins.

Our customers are building many of these themselves, using third-party and open-source tools alongside our equipment. We’re also partnering with a number of companies to test and document proven integrations. Our new VP of Business Development, Ali Hedayati, has his hands full with all of these relationships and others.

Whatever the final result, there’s no doubt that user experience is a ripe field for innovation, and that it’s transforming many parts of an organization far beyond simple incident detection.

After a whirlwind of activity in late 2006, I get a brief respite to reflect on things before hitting the road again next week. We had a great 2006, far beyond our wildest expectations, and it feels like we’re spending less and less time explaining what Real User Monitoring is and more time understanding how it’s going to change the way a particular industry does business.

One example of this came across in the past month. While we’ve got customers in all kinds of industry verticals, from healthcare to finance to entertainment to e-commerce, one of the places we’re particularly strong is in the Software-As-A-Service (SaaS) sector. This class of web applications offer software functionality with a hosted model. SaaS heavyweights include companies like Salesforce.com (salesforce automation and CRM), RightNow Technologies (CRM and customer support), Taleo (Human resources and recruiting), and ADP (HR and payroll.)

Most of these companies collect revenue through “seats” — that is, the number of users subscribed to the application. It’s a lot harder to regulate account abuse in a SaaS model than it is with a software license. When I sell a copy of some software, I can include a license key and some form of online verification, which prevents sharing. And that license isn’t very portable: The overhead of uninstalling it on one machine and installing it on another is usually overwhelming. But as a SaaS vendor, I can’t enforce licensing in this manner when one of my main value propositions is that there’s nothing to install!

To make matters worse, another key advantage of SaaS is the portability and mobility of the application. Customers can use Salesforce.com from home, or work, or anyone else’s machine. So how does Salesforce know they’re not sharing user accounts?

It turns out that this is a major issue for many of our SaaS customers. Some real estate offices have a single account with a SaaS provider, but an entire office uses that account. Recruiters give out their password to people in other countries. Some salespeople adopt a “timesharing” approach to hosted applications. It’s astonishing how creative end-users can get when it comes to saving $40 or $50 a month!

As you might imagine, Real User Montioring is a great tool for detecting and proving account abuse. We’ve recently added all kinds of geographic lookup and service provider tracking capabilities to the latest release of TrueSight, and some of our recent customers — who initially came to us for incident detection and service level reporting — are using us for account fraud detection. It’s easy to win an argument with an end-user when you can prove that he apparently got to work in Texas, had lunch in Bangalore, and ended the day in France. And I’m not making that one up. Sounds like my travel schedule.

Of course, some of this stuff can be detected within the application itself — most SaaS tools won’t allow multiple logins from the same account. But few of the SaaS providers have properly instrumented this, much less tied it back to an account team who can then call the customer and sell them a few more account licenses. And with all of the sneaky tricks people play in order to circumvent additional licenses, the forensic information you get from Real User Monitoring is invaluable. For example:

• You might get several logins that look like they’re from the same IP — but have different X-Forwarded-For headers because they’re really different machines behind a proxy
• You might get different user-agents within a session as people switch browsers
• The TCP round-trip-time (a measure of delay across the Internet) may vary wildly, indicating different latencies behind a hop on the net

One of the best things about working with a bunch of smart customers is that they’re always pushing the envelope. In our case, Coradiant’s Session Networking technology — the stuff that powers our TrueSight boxes — has all sorts of interesting and creative uses we hadn’t thought of when we set out to make a latency monitoring appliance back in 2003. And we had no idea we were going to build an appliance when we started monitoring user experience on the sites we were operating as an MSP back in 2000.

The holiday seasons are always a time to reflect, and it’s been a fascinating few years discovering all of the unexpected ways this technology can be put to good use throughout the Internet world. Here’s to an equally fascinating 2007!