HP: A Trusted Source for Advice about Big Data?
July 9, 2015
Remember that Hewlett Packard bought Autonomy. As part of that process, the company had access to data, Big Data. There were Autonomy financials; they were documents from various analysts and experts; there were internal analyses. The company’s Board of Directors and the senior management of the HP organization decided to purchase Autonomy for $11 billion in October 2011. I assume that HP worked through these data in a methodical, thorough manner, emulating the type of pre-implosion interest in detail that made Arthur Anderson a successful outfit until the era of the 2001 Enron short cut and the firm’s implosion. A failure to deal with data took out Anderson, and I harbor a suspicion that HP’s inability to analyze the Autonomy data has been an early warning of issues at HP.
I was lugging my mental baggage with me when I read “Six Signs That Your big Data Expert, Isn’t?” I worked through the points cited in the write up which appeared in the HP Big Data Blog. Let me highlight three of these items and urge you, gentle reader, to check out the article for the six pack of points. I do not recommend drinking a six pack when you peruse the source because the points might seem quite like the statements in Dr. Benjamin Spock’s book on child rearing.
Item 2 from the list of six: “They [your Big Data experts] “talk about technology, rather than the business.” Wow, this hit a chord with me when I considered HP’s spending $11 billion and then writing off $7 or $8 billion, blaming Autonomy for tricking Hewlett Packard. My thought was, “Maybe HP is the ideal case study to be cited when pointing out that someone is focusing on the wrong thing. For example, Autonomy’s “black box” approach is nifty, but it has been in the market since 1995-1996. The system requires care and feeding, and it can be a demanding task mistress to set up, configure, optimize, and maintain. For a buyer not to examine the “Big Data” relevant to 15 years of business history strikes me as important and basic step in the acquisition process. Did HP talk about the Autonomy business, or did HP get tangled in the Digital Reasoning Engine, the Integrated Data Operating Layers, patents, Bayesian-LaPlacian-Markovian methods?
Item 4 from the list of six: “They [your Big Data experts] talk about conclusions rather than correlations.” Autonomy, as I reported in the first three editions of the late, lamented Enterprise Search Report, grew its revenue through a series of savvy acquisitions. The cost and sluggishness of enterprise software sales related to IDOL needed some vitamin supplements. Dr. Mike Lynch and his capable management team built the Autonomy revenue base by nosing into video, content management, and fraud detection. IDOL was the “brand,” and the revenue flowed from the sale of a diverse line up of products and services. My hypothesis is that the HP acquisition team looked at the hundreds of millions in Autonomy revenue and concluded, “Hey, this is a no brainer for us. We can sell much more of this IDOL thing. Our marketing is much more effective than that of the wonks in Cambridge. Our HP sales force is more capable than the peddlers Autonomy has on the street.” HP then jumped to the conclusion that it could take $700 or $800 million in existing revenue and push it into the stratosphere. Well, how is that working out? Again, I would suggest that the case to reference in this Item 4 is HP itself.
Item 6 from the list of six: “They [your Big Data experts] talk about data quality, rather than data validity.” This is an interesting item. In the land of databases, the meaning of data quality is often conflated with consistency; that is, ingesting the data does not generate exceptions during processing. An exception is a record which the content processing system kicks out as malformed. The notion of data validity means that the data that makes it into a database is accurate by some agreed upon yardstick. Databases can be filled with misinformation, disinformation, and reformed information like a flood of messages from Mr. Putin’s social media campaigns. HP may have accepted estimates from Qatalyst Partners, its own in house team, and from third party due diligence firms. HP’s senior management used these data, which I assume were neither too little nor too big to shore up their decision to buy Autonomy for $11 billion. As HP learned, data, whether meaty or scrawny, may be secondary to the reasoning process applied to the information. Well, HP demonstrated that it made a slip up in its understanding of Autonomy. I would have liked to see this point include a reference to HP’s Autonomy experience.
Net net: HP is pitching advice related to Big Data. That’s okay, but I find that a company which appears to have struggled with Big Data related to the Autonomy acquisition may not be the best, most objective, and most reliable source of advice.
Talk is easy. Performance is difficult. HP is mired in a break up plan. The company has not convinced me that it is able to deal with Big Data. Verbal assurance are one thing; top line performance and profits, happy customers, and wide adoption of Autonomy technology are another.
The other three points can be related to Autonomy. I will leave it to you, gentle reader, to map HP’s adult-sounding advice to HP’s actual use of Big Data. As the HP blog’s cartoon says, “Well, maybe.”
Stephen E Arnold, July 9, 2015
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
