So Salesforce.com yesterday acquired Tableau for $15.7 billion in stock. This follows on the heels of two other notable acquisitions: Google’s purchase of Looker (for $2.6 billion) and Logi Analytics’ acquisition of Zoomdata. My take? It’s all about blockchain. I kid.

Seriously: I kid.

Salesforce-Tableau isn’t strictly a “BI” acquisition, per se, but I go back to 2007, when the top BI pure-plays fell (almost sequentially) like dominos: Oracle bought Hyperion to kick things off; seven months later, SAP responded by acquiring Business Objects. IBM capped the cycle in November of 2007 when it purchased Cognos. All three were the dominant pure-play vendors in the BI/analytics space. All three had fattened themselves on (much smaller) BI and analytics pure plays in the 4-5 years prior to their own acquisitions. And, in a stretch of just over half a year, all three got gobbled up by larger players. Their acquirers were symmetrical competitors in several different markets, with one market (enterprise applications, or what used to be called ERP) emerging as a particular locus of battle, if not a common thread.

So what’s the common thread today? The truth is that I haven’t the foggiest [censored] idea. I can point to certain signs that I believe are indicators of possible trends—e.g., the endangered independent or pure play software vendor? Innovation-by-acquisition? A drive to shore up shortcomings in BI/analytics/data management products and services? The herd instinct that runs rampant whenever some imperceptible shift in the Zeitgeist transforms an industry?—but I won’t pretend that my insights are any more valid than those of, say, a haruspex.

My friend Mark Madsen pointed out that Google had made several announcements prior to its acquisition of Looker that, in hindsight, seem to dovetail with Microsoft’s classic SQL Server-based BI strategy: You’ve got Big Query BI Engine, a server component that’s akin to SQL Server Analysis Services, for one thing. Plus, Looker gives Google self-service discovery environment that’s at least notionally analogous to PowerBI. (Okay, Mark admits that this last is a stretch.) Google also boasts not one but two SSIS-like cloud services: Big Query DTS and Cloud Data Fusion. (There’s also that product name, “Big Query DTS.” It’s almost as if Google has Microsoft’s pre-Azure SQL Server BI stack on its brain.) So is Salesforce.com’s acquisition of Tableau a response or reaction to Google’s aggressive push to flesh out its BI, analytics, and data management portfolio? What about LogiAnalytics’ acquisition of ZoomData?

To a degree, yes, obviously they are—at least in the sense that all of these vendors swim in the same big ocean. Is there something more to it than this? I. just. don’t. know.

That said, I think Tony Baer has a great take on what’s going on: it’s all about multi-cloud. Tony’s is a great take. I just don’t believe it’s the only take. I don’t think he does, either.

Another friend, a former analyst turned product visionary with a large analytics vendor, has a slightly different take. This person, too, sees the recent consolidation cycle as being driven in part by vendor positioning for the cloud. “The old stack integration model was still best-of-breed and the vendor had to tie all of the parts together. Now the parts are hidden under a service layer. They must still be integrated or the seams will show to the users and that is bad,” this person says. “So everybody with a cloud needs a cloud be a stack story, and a place for the data to come from. If you look at Google they did a lot to work out service integrations with many different application providers, [this is] expensive work that used to be amortized across many customers by one company: Simba. Deep monopoly pockets enable you to do this and do it at scale. I believe Google claims integration with over 1,000 services now.”

Industry seer Wayne Eckerson views this as a missed opportunity for at least one vendor—Amazon. He, too, sees the Looker and Tableau acquisitions as akin to a game of chess in the cloud. “The Tableau acquisition is definitely a cloud vendor solidifying its position -- but it's odd that Amazon didn't purchase it, given that its new CEO came from Amazon and Amazon needs better BI tools. Definitely a missed opportunity for Amazon,” he writes. “And I'm not sure what Salesforce will really do with Tableau since it's an app vendor. But someone closer to Salesforce would know more about their current strategy and how this fits.”

Maybe it’s time for Amazon to take another (albeit longer) look at Yellowfin BI?

It’s possible to see all of these moves as related, to a degree, to the same phenomenon that both Wayne and my analyst friend-turned-corporate-talking-head describe. That being said, with respect to Salesforce’s acquisition of Tableau in particular—i.e., considered separately from other acquisition activity in the space—I think there’s another obvious explanation, too: both companies have massive corporate overlap—i.e., many of the same corporate customers use both Salesforce and Tableau; most people doing any kind of analysis in Salesforce use Tableau, albeit perhaps in combination with Excel, Alteryx, and a few other tools.

Salesforce is just making it official.

And, in the process, gaining bragging rights vis-à-vis direct and indirect competitors alike.

*

These bragging rights will come at a cost, however. If I’m in a non-technical position with Tableau or Looker, I’m at once excited and terrified. The economic impact of all of these acquisitions is probably going to be, on the whole, a net negative for a fair proportion of employees with these companies. The troubling upshot is that many people (particularly those in non-engineering positions) are going to lose their jobs. All mergers and/or acquisitions entail labor-force reconciliation or rationalization of some kind. In most cases, this will take time. But the jobs most likely to be “rationalized” (i.e., eliminated as duplicates) are of the bureaucratic-professional type. As the reliably oracular Mark Madsen soberly puts it, “the combined Salesforce/Tableau and Google/Looker will not need an HR department, or a finance department, or an accounting department, or a security department, or an IT department, etc. All of these are ancillary functions of the company outside of core product development and product marketing, and likely many of those marketing jobs will be gutted.”

Another casualty will likely be the character of Tableau’s utterly unique fanb—er, customer base. For Salesforce customers, this is a win-win—the M&A equivalent of formally putting a ring on the finger of the sweetheart to whom one had informally plighted one’s troth.

Tableau’s combination of best-in-class data visualization and self-service data-driven discovery features could make even Rudy Giuliani—no friend to rigorous analysis, let alone empirical reality, statistical numeracy, discovery of hidden and/or occult influences, etc.—learn to stop worrying and love analytics. To the extent that Salesforce’s existing analytical or data visualization feature set is impoverished—and I cannot speak to this, as I haven’t looked closely at Salesforce in several years—Tableau will make it considerably less impoverished. 

To the extent that Salesforce customers are stymied by messy, incomplete, or inconsistent data, Tableau will help with that, too. (Not that Tableau itself has an especially powerful data management or data quality feature set; its shortcomings in this regard are things the company has struggled for years to address, actually. But data discovery is a useful paradigm for quickly diagnosing data quality problems, too. Fixing them is of course another matter.) The acquisition is a boon for Salesforce customers—not unlike being paired with Jeopardy! champ James Holzhaeur on pub trivia night. There’s a sense, too, in which Tableau might not present all that steep of a learning curve for Salesforce users—or analysts, at least. Advanced users probably already use Tableau (or something like it) against their Salesforce data today. What’s more, there’s a huge overlap between Salesforce’s and Tableau’s customer bases: They’re both in oodles of the same accounts. So they’ve got them going for them, which is nice.

My biggest question has to do with what this will mean for Tableau’s users.

If you’ve never been to one of Tableau’s customer conferences, you can’t possibly imagine how important Tableau is to the people who use it. You might be forgiven for feeling as if—in the midst of one of Tableau chairman Christian Chabot’s keynote speeches, for example—you’re voluntarily standing in a queue for grape-flavored Flavor Aid at Jonestown. I don’t mean this as a knock against Tableau-phytes, either. Self-service BI, and Tableau in particular, addressed the business person’s unmet demand for agency—i.e., for self-determination and autonomy. Business people, aspiring analysts, especially, wanted to take some control back from an IT octopus that had become obsessed with controlling access to IT resources—and which gave priority to restriction, as distinct to accessibility. Tableau, more, perhaps, than any other tool, helped them do that. It was at once tool and weapon. The people who love it are, in my experience, even more fana—er, loyal than the most ardent of Apple acolytes.

Is today the day the music died for Tableau-phytes? Certainly, they’re apprehensive. Salesforce is an outside behemoth. It, too, has its culture of loyal customers, but, in my experience, the average Salesforce user is less totally … committed than the average Tableau-phyte. To put it differently: Maroon 5 has its fans; Vampire Weekend has its obsessives. It’s difficult to imagine the two of them sharing the same stage, let alone having simpatico audiences. Something will have to give. It makes me think of Neil Diamond’s going on stage before Bob Dylan at the Band’s Last Waltz concert: “Follow that,” Diamond is alleged to have teased Dylan when he came off stage. “What do I have to do, go on stage and fall asleep?” Dylan is said to have riposted. Marc Benioff is exciting, but he ain’t Christian Chabot.

**

Something else will have to give, too. By this I mean to say that it’s necessary to distinguish between Tableau’s users and Tableau’s customers, i.e., Tableau’s corporate buyers.

To a degree, the term “corporate buyers” is still synonymous with “enterprise IT organizations,” although this identification is less absolute-ish than it used to be. (The do-it-yourself or bring-your-own ethic could just as easily be called the charge-it-yourself ethic.) Rank-and-file users were hip to Tableau’s alterity long before enterprise customers ever glommed on to Tableau as, you know, a thing. It is thanks to an insurgency on the part of rank-and-file users that Tableau established its enterprise beachhead in the first place. But that’s also the rub.

Tableau’s rank-and-file users have different needs and priorities than its enterprise customers. The people cheering and ooooh-ing and ahhhh-ing at Tableau customer conferences? Those were Tableau users, not corporate buyers. Priorities such as data lineage, metadata management, and consistent master data definitions—to say nothing of a single version of the truth (however hallucinatory that vision might be)—are of much less importance to rank-and-file Tableau users than they are to data management practitioners. Tableau established its beachhead in the enterprise by catering to the needs of rank-and-file users; thereafter, it shifted course to address the needs and priorities of enterprise customers—and, particularly, of data management practitioners. To Tableau’s credit, this focus has borne empirical fruit. 

Just a few months ago, for example, Tableau announced its “Data Management Add-on for Tableau Server.” Think of this last as a self-service take on responsible—or, rather, governable—data preparation/integration. One point of emphasis with Tableau’s Data Management Add-on for Tableau Server is repeatability: via its Tableau Conductor and Tableau Prep facilities, Tableau now permits users to instantiate data flows as scheduled, repeatable processes to Tableau Server. From an IT or data management perspective, this is the bee’s knees: repeatability and schedule-ability—along with, perhaps most important, traceability—are core motivating priorities for IT buyers. If for some a Tableau Conductor data flow does not complete as expected, IT would prefer that it log alerts, exceptions, or other relevant contextual information. If a Tableau Conductor process results in the movement, manipulation, and/or transformation of data, IT would prefer that it generate a metadata record of these changes (its “lineage”)—e.g., who changed what and how—along with any additional metadata that describes any newly created objects, entities, etc.

Tableau’s Data Management Add-on for Tableau Server does just this.

With this in mind, Salesforce’s acquisition of Tableau invites an irresistible question: what will become of Tableau Server? The good news is that this is a less irresistible—or, at any rate, less pressing—question today than it was, say, five years ago. Today, and irrespective of its ongoing centrality in or for Tableau’s architecture, Tableau Server is in some sense a relict product: a vestige of the before-cloud earlies. Tableau was already looking beyond Tableau Server. (Which is to say: even though it still has its place, Tableaus Server is no longer the orbital center of the Tableau paradigm. This problem isn’t unique to Tableau, either. The shift to cloud-native architecture will force most BI and analytics players to radically revamp their product architectures.) A little more than half a decade ago, Tableau introduced its cloud offering—basically, a version of Tableau Server running in a cloud context. Like almost every cloud offering from almost every BI/analytics player, this first go-around wasn’t in any sense a slam dunk. Since then, however, Tableau has introduced a native cloud service, Tableau Online, that is a little bit closer to the proverbial (Kawhi Leonard) slam dunk.

All of this is to say that, while Tableau’s classic client-server-oriented paradigm isn’t a good fit for Salesforce, its native cloud paradigm likely is. It’s also to say, first, that a staggering number of Tableau customers (some of them Salesforce subscribers) run Tableau as part of a conventional, client-server-ish deployment model and, second, that that the needs and priorities of these customers don’t necessarily align with those of Salesforce. To the extent that existing customers would prefer a Tableau that (a) continues to focus on core data management priorities, (b) sustains its role as a best-in-class provider of data visualization and discovery features, and (c) isn’t tightly coupled (or explicitly tethered) to a single provider’s application stack, you might could say that the needs and priorities of Tableau’s customer base are orthogonal to those of Salesforce subscriber base. How do you square that circle?

In all likelihood, you don’t. You’re back to something’s-gotta-give. After all, as Wayne drily notes, the trope of “app vendors purchasing BI vendors has never been very successful.”

***

As for where we go from here—I’ve no idea. One thought is that other significant pure plays, perhaps in tangential markets (e.g., Alteryx) might either be acquired or begin courting suitors to acquire them. This last is something that always seems to happen whenever an industry acquisition cycle kicks off. Maybe it’s Alteryx or Trifacta or Paxata or SiSense or some other similar vendor. Maybe the metadata catalog space is due for a round of consolidation.

The lesson with the consolidation cycle of 2007 was that—in retrospect—it seemed obvious that once one of the Gang of Three got acquired, the others would (have to) get acquired, too. 

A similar phenomenon happened in the data warehouse (DW) appliance segment, which underwent a consolidation cycle from about 2011 through 2013. One by one, each of the DW appliance pure-plays tumbled like so many dominoes. And in this case I am able to say that one of the last to fall, the former ParAccel, was anxious to get itself acquired. Why was that? Part of it had to do with pressure to make its VC investors whole, of course. And by 2013, ParAccel wasn’t exactly basking in its status as the last standing DW pure-play: behind the scenes its investors were getting antsy. There was the distinct sense that it was time to sell.

But what drives these cycles in the first place? Why do these otherwise loosely connected (or coupled) dominos all start to fall at once? Again, I. just. don’t. know. No one else does, either.

I’ll close with a prediction—actually, the recycling of what we in la cosa nostra regard as that hoariest of chestnuts: Informatica will soon and inevitably be acquired. In the fullness of time.

You read it here first.