Blog: AI expansion, IT commoditization
At the dawn of CRM there were 5 or 6 different application stovepipes that have since been consolidated. There were once separate products for inbound customer service and outbound telemarketing though we hardly ever talk about the distinction any more. There was also a help desk app for serving internal employee customers of big companies that needed help with tech issues. But today we think primarily about sales, marketing and service with a few additions like ecommerce and analytics.
Marketing was the last of the apps to crystalize, in part because no one knew what a marketing app should do, believe it or not. Early versions were basically accounting systems that tracked marketing spend so that we’d at least know how our money was being wasted.
When we finally got around to doing marketing, we discovered that it wasn’t about accounting at all, in fact, we found out that it wasn’t even a single app. Marketing sprawled into numerous flavors. There was marketing for email, social media, data gathering and statistics, outbound, inbound, journey mapping and a lot more. Today the drumbeat continues with chatbots and omnichannel marketing and it’s difficult to know where it ends though we know that trees don’t grow to the moon.
It’s becoming clear now that marketing was not an anomaly, not the only thing that could sprawl and the broad category of BI or analytics looks to be the next part for the front office whose footprint can’t be determined yet. Earlier this week Salesforce offered to buy Tableau for $15.7 billion in an all stock deal that was roughly 50 percent above market cap, and Tableau said sure. The transaction will wrap by end of October in plenty of time for Dreamforce and the first real discussion of how the two companies fit together.
Salesforce began its analytics sojourn with Salesforce Wave, later called Salesforce Analytics Cloud which primarily provided BI for sales. All of that rolled into Salesforce Einstein which is used for making next best recommendations and other reporting tasks. There were so many parts to this evolution that I may have dropped one or got a name wrong.
The article covering the acquisition in Tech Crunchlisted Tableau as a data visualization company. Also, Bloomberg’s article, “Salesforce Patches Up a ‘Flop’ With $15 Billion Bet on Tableau”gets some stuff not quite right and it’s worth trying to clarify especially where Bloomberg repeatedly refers to the category as Business Intelligence.
The differences are subtle and some might say why bother splitting hairs. Business Intelligence (BI) tries to deeply analyze company and customer data to help businesses figure out what will likely happen based on past history. Which customers are in danger of leaving a vendor, which deals are most likely to close before the end of a quarter, and based on our history with customers what is our best next offer or suggestion to a customer in a particular situation?
ML studies data and builds models based on patterns found in the data that drive algorithms. It runs on inference rather than explicit instructions. The classic example is the retail discovery that placing disposable diapers near the beer cooler sells more beer in a convenience store. Apparently, new fathers run out of the house to get a resupply of diapers and snag a six pack on the way out the door. Learning models discovered the relationship not surveys or any active merchandising.
When machines approach thinking like humans, the term artificial intelligence might come into play. Understanding human speech is an example, so is the ability to play a complex or strategic game such as Jeopardy (thanks, Mr. Watson) or chess (Deep Blue); autonomous cars are an example nearly everyone has a notion of.
The commoditization of IT
All of this comes into sharp relief when we begin considering the commoditization of IT. Information processing was a disruptive innovation in the late 1940s that was given a turbo boost with a patent granted in 1959 for the silicon chip on which all computer componentry, especially CPUs, are based today. But the chip was a disruptive innovation capable of commoditization and so was cloud computing which is to some minds (okay, mine) the ultimate commoditization of IT.
Commoditization is important because it signals declining profitability (commodities are cheap) which results in shrinking an industry’s leadership into an oligopoly or even a monopoly. IT is commoditizing and BI, ML, and AI are accelerating the process.
Another part of commoditization is the deleterious effect it has on jobs. Often called automation, it reduces headcount in favor of machines. But jobs in IT won’t evaporate so much as they won’t come into being with this wave of automation. Companies like Salesforce, Oracle, and Microsoft are working to build autonomous customer facing systems just as sure as Google and Tesla are building autonomous cars.
All of the big enterprise software companies are working to build systems of engagement that literally participate with customers to figure out and deliver goods and services with the lowest costs and highest profitability. That said, when the Bloomberg piece said of the acquisition,
The deal means that Salesforce, whose key to success has been a dominant position in customer relations, is becoming more of a general-purpose information technology company — following Oracle’s transformation from database giant to IT conglomerate over the last 20 years, when it snapped up new products and customers.
Business intelligence “is not Salesforce’s core competency and there is much Tableau does that doesn’t pertain to the CRM world, making the fit slightly imperfect,” Chrane [an analyst at Sanford C. Bernstein & Co.] wrote Monday in a note.
It beggars belief.
No one can have a dominant position in CRM unless they have a clear conception of what customers want now and in the future AND unless they can provide it at the lowest possible cost, two things that automation is essential for.
Your customer wants this pair of shoes in a size 12.5, red. They’re not in the back room. Where in the chain of stores is the nearest pair? How long will it take to get them? If they aren’t in the chain, are they in the warehouse? How long? When does the customer want them?
This real world problem (yes, I wear red shoes in that size) isn’t something that gets spelled out and initiated by a sales clerk. It’s something customer facing systems, systems of engagement, do routinely these days. So when we say that a core competency is customer relations but that business intelligence is not a core competency, I’d argue that not BI but AI is what needs to be core for almost any business moving on.
The descriptions of Tableau as a data visualization company is accurate-ish. It was founded in 2003 before AI had much sway but it bought Empirical Systems in 2018 for its AI capabilities. As a visualization company, Tableau can do amazing and interesting thigs with the output of AI like mapping data points to reveal what the shape of information solutions look like. This engages the right side of the brain while the raw numbers engage the left thus giving the user a more wholistic view of business or any problem at hand and helping foster solutions.
My two bits
This isn’t your grandfather’s BI world anymore and it makes good sense to traverse the path from BI to ML to AI as quickly as one can. Visualization is a part of that path because it simplifies an arcane academic pursuit, data science, and turns it into a business tool that anyone can use. Is $15.7 billion a lot of money? Let’s just say a much smaller amount would make for a great weekend in Vegas. Is it too much for Salesforce to spend on Tableau? Better question but who knows? It’s not the cost of the investment but the expected return that counts. Right now, that’s up in the air so prognosticators are having a field day, but it fits with a progression of taking on increasingly difficult customer facing problems and the return on that is also incalculable.