Tealeaf, IBM and Warehouse Technology Stack

Acquisitions are, by far, the most significant milestones in a technology landscape. Sure, new technologies and brand new vendors can be transformative. Hadoop is having a profound impact on the high-end data analytics warehousing ecosystem. Radian6 was transformative in social media measurement. In an industry like ours, however, acquisitions tend to rule. Think about the acquisition of Urchin by Google; Sane by Unica, IBM of Unica (and Coremetrics, Netezza, etc.), WebSideStory of Visual Sciences, Omniture of that combined company, and Adobe of Omniture (and Offermatica, TouchClarity, Efficient Frontier, etc.). We've seen a continuous ascent up the food chain as small analytics vendors were acquired or grew into mid-size companies who were then acquired by behemoths. Analytics is now a cornerstone of the growth strategy of the world's largest enterprise technology companies.

All of which is just a prelude to thinking about the IBM acquisition of Tealeaf this past week. Tealeaf isn't just the leader of the Customer Experience Management (CEM) space, they are almost it's de facto owner. For true enterprise CEM analytics, Tealeaf is pretty much the only game in town.

Tealeaf, as a company, is a natural fit for IBM. Not only is IBM building up a comprehensive suite of digital measurement and analytics tools (a suite that's strikingly different in shape than Adobe's), they play heavily in areas where Tealeaf is dominant: call-center and customer operations. A great many of Tealeaf's core clientele aren't really analytics clients; they are customer operations and call center clients who use Tealeaf primarily for web-related customer support problems. Unlike most analytics vendors (such as Adobe), this class of enterprise problem is meat-and-potatoes to IBM.

So IBM has a natural path to value in terms of Tealeaf's core business. That's always a good thing when it comes to acquisitions. But I'm more interested in whether or not IBM has a deeper interest in Tealeaf and how Tealeaf might be a part of the broader IBM digital suite.

Semphonic is a relatively new (just since the beginning of this year) Tealeaf partner, but I've long believed that Tealeaf was one of the most under-utilized tools in the digital analytics space. Many of our clients have Tealeaf tucked away in their call-center operations area and get little or no use from it analytically. It's a shame, though there are some reasons why Tealeaf often get siloed.

What's unique about Tealeaf is also what makes it challenging. Tealeaf is a sniffer - collecting data from the HTTP stream as it passes to and from the user to the client servers. Unlike some other sniffers, however, Tealeaf doesn't really weed, filter, and structure the data as it passes through. Tealeaf saves pretty much all of it. This makes Tealeaf expensive, but it also drives significant value. From a customer service standpoint, saving everything is like never having to say you're sorry. You don't know what data is going to be significant, so you pretty much have to save everything.

Analytically, of course, that's not always true. You CAN make pretty shrewd guesses about what's going to be significant. That's what tag requirements and design are all about. But as I discussed in last week's Webinar with Tealium on Tag Management Systems, that process of building requirements is non-trivial and error-prone.

So there's real virtue in a "save-everything" kind of approach. It's the same virtue I lauded in Celebrus. What's more, while the Tealeaf approach is technology expensive, it's people cheap. You don't need two or three web analytics consultants full-time onsite building tagging requirements when Tealeaf is your data collection mechanism. It's also zero impact: no page weight, no page changes, no site risk.

And here's another consideration. Systems like Tealeaf support a fundamentally different type of interface into the data than traditional Web analytics systems (or BI or Statistical tools for that matter). When you're collecting HTTP data streams (which are inherently unstructured), search becomes the primary query language. Users of traditional systems tend to be skeptical of search as a primary interface into the data (I know I am), but take a look at a system like Splunk and you can see that there is real power in the approach.

So I'm wondering if IBM isn't seeing Tealeaf as the potential data infrastructure piece for their entire digital suite. To make that happen, they'd have to provide tools for building structured views of the Tealeaf data (something Tealeaf already does). A fair number of tools in the suite wouldn't work otherwise. It's not that difficult a task to extend Tealeaf's existing capabilities, however, and once accomplished, IBM would have a unified data collection piece that completely bypassed the whole world of tagging and tag management. Yes, there are some drawbacks to this. Tagging handles client-side tracking of interactions that simply aren't capturable server side. Still, with a Tealeaf-based infrastructure, IBM would be able to support a wide range of analytics, call center and operational needs without any pre-planning or tagging. They would be able to source multiple tools all from a single real-time and exhaustive collection piece. They would be able to support both structured and un-structured access to the data. That might be pretty compelling.

It wouldn't be a solution for everyone, but for IBM's core enterprise-class clients, you can see that it might have real advantages as a comprehensive real-time analytics and warehousing infrastructure.  

Which, as it happens, is a topic I've been thinking quite a bit about lately. Here at Semphonic we've been doing an ever increasing amount of work around the analytics warehouse and a real-time technology stack is one of the issues we keeping running into. I plan to talk more about some of the tools (most of which are a bit obscure) that potentially fit inside that technology stack in upcoming posts.

 

[Notes: Last year at X Change, we did the first Non-Profit Challenge. A day of deep-dive analytics by X Change participants into the analytics problems of two large non-profits. We recently published the results of that work in the Non-Profit Guide to Using Analytics Whitepaper. Not only can you download that whitepaper - a truly collaborative effort - if you're a non-profit, I strongly encourage you to tune into this week's webinar with Emily Fisher of Oceana, Linda Shum of United Way, and our own Phil Kemelor as they discuss the work, the whitepaper, and the use of digital measurement in the non-profit space.

And speaking of the data warehousing technology stack, I'm doing a webinar the following week with one our partners, iJento, with more on customer journey tracking. We've partnered with iJento because they provide a robust traditional (SQL-Server) database platform that can leverage our Two-Tiered segmentation data model (and collection mechanisms like Celebrus). iJento probably isn't the right solution for the 1% with extreme digital data volumes. But for companies with large but manageable digital data volumes, they provide a much less risky technology stack (and both SaaS and on-premise models) that can deliver deep access to customer digital data on top of a robust platform with a lots of support, great software, and tremendous flexibility.

Finally, I'll be leaving soon for Germany (I have some client engagements there prior to X Change). If you're EU-based and would like to setup meetings in Stuttgart, Berlin, Paris or London, drop me a line. I'd love to chat!].

Comments

[Michael Dickey]

Pion from Cloudmeter (http://www.cloudmeter.com) is similar to Tealeaf in that it uses a passive packet sniffer to capture clickstream data with visitor session replay capabilities. Pion allows you to "store everything" as well as filter/clean/aggregate your data in real-time. Unlike Tealeaf, it also provides a very graphical, web-based interface to define real-time ETL processing & manage third party integrations with any popular database or BI tool.