When the music stops, and the feel good from having the Beach Boys drill the Salesforce Wave analytics message into your brain wears off, I am left wondering what the buyer will really think. Wave is problematic on multiple levels and while I am penning this, I have a slew of unanswered questions.
Analytics is a minefield for any vendor. Positioning has to be pristine and to that extent, the company did a masterful job. Having executives from Coca Cola Germany waxing lyrical in un-Germanic fashion, having GE Finance talking about 'this changes everything' and the same at Honeywell speaks to several themes.
These are three of the most well known brands on the planet. In Europe. In SAP's back yard. Couple that with the brilliant one liner:
We don't need a new version, we need a new vision
And the message is clear: 'SAP? We're comin' after you and this time, we really mean it.' But it is when you look at the manner in which it was announced and the lack of clarity on key points that things unravel. And this year - they don't get a pass.
For the moment, I am going to restrict myself to an overview of three topics in the analytics story: pricing, technology and the ISV partner issue but first let's look at how Wave has been brought to us.
First there was the Benioff leak. Next there was a roundtable which was more like a village hall meeting during which the key presenters could not answer any of my questions. Next came the Dreamforce keynote where CEO Marc Benioff wheeled out a more flashy discussion on Wave. Then we got to meet some of the technical team who gave us some answers but left us head scratching - again. Today, there is the Wave keynote. Bizarre? You bet.
In my earlier piece I referenced the 'builder' pricing at $250/user/month and 'explorer' pricing at $125/user/month. This doesn't make sense. Salesforce is anticipating a distribution of 1 builder to 9-10 explorers. Let's look at what this means at various break points:
- 5 person - $750/month
- 10 person - $1,375/month
- 25 person - $3,375 per month
- 50 person - $8,125/month
- 100 person - $13,750/month
I've made a few assumptions in here and these are pre-discount pricing models but you get the idea.
Keith Bigelow, GM & SVP, Analytics Cloud Salesforce believes that is price competitive with the likes of Tableau. That's not true. Here is Tableau's pricing.
- The personal desktop version of Tableau runs $999/user (one time license fee) plus maintenance.
- The pro edition, which can be equates to 'builder' runs $1,999 per user plus maintenance.
- The hosted version runs $500 per user per annum.
In a very small 5 person business I can get productive for $2,500 per annum or $208/month. In reality, I can run a small business for something around $2,000 per annum and distribute interactive dashboards to pretty much anyone for free. That doesn't seem to be the case with Wave.
In response, Salesforce sees itself as pitching against SAP and Oracle. That's the wrong positioning. Neither of those companies provide equivalent functionality at the lower end scale which can compete on price/functionality with Tableau as a vizualisation tool which is where I see Wave's direct and immediate competitive positioning. Similarly, Tableau is aiming to hit more $100,000 deals in its sales strategy, yet many of its customers are well established brands. And make no mistake, Tableau is the big dog in this space - today.
In return, some Salesforce executives believe that the company is really aiming Wave at the high end of its customer spectrum. If true then those folk will have little difficulty in 'getting' the pricing as currently outlined but even there I think Salesforce has this wrong. You cannot have disparate pricing for different parts of the market - you upset one or another, or, worst case scenario, both. At the high end, we will be talking significant discounting and enterprise pricing - in short, lump sum deals that bear no relation to the price book for applications that will run on dedicated instances of Salesforce/Wave.
I reminded Salesforce executives that we've seen this pricing faux pas before with Chatter which started out at $50/use/month and got nowhere. Now it is free though you pay for Communities. Wave will go the same way. Why?
The 'mobile first' strategy is superb differentiated positioning. It knocks Tableau straight off the rails, even with its responsive design approach. Having seen the ability to quickly set parameters around what you want to discover and filter information, I can see a burgeoning audience among sales staff who are on the road. But is that worth $125/user/month? Not for this buyer.
The consensus is that Salesforce will move quickly to fix any pricing issues and of that I have no doubt. What disappoints is that we appear to have a case where history is repeating itself.
So far, I have had nothing to say about any of Salesforce's past offerings from a technology standpoint. They work for the purpose for which they are intended even if that underpinning is now a very old, highly customized Oracle database. But for operational analytics, technology matters. Why?
It is very clear that Salesforce has optimized Wave for Salesforce core data. To that extent, it's mish mash of open source based databases and technologies like Riak, Cassandra and HBase (note - not Oracle) is smart and works well. The company says it is running on commodity hardware but doesn't say what compute power is needed for any given data set. In short - we don't know what it takes to run at scale and the extent to which that impacts pricing downstream.
Compare that with SAP. As regular readers will know, there has been an ongoing criticism of SAP for offering a highly scalable, super fast but insanely expensive solution in SAP HANA. That criticism is not going away but I would argue that it is far from clear whether Salesforce can even get on nodding terms technology wise and of course the pricing issue remains open for Salesforce as well.
Not a single Salesforce technical executive could provide me with any meaningful benchmarks. Neither could they say the extent to which Salesforce has tested for individual pod scale out or customer load. That is a worry. Why?
Optimizing for Salesforce data is one thing but when you carry the promise of both Force.com data and external data sources then you run into problems very quickly. Here's why.
Salesforce has no native extract, transform and load layer except for its own data and there is no obvious way to optimize for those other data sources. It is currently relying upon Informatica to provide that tooling. Fine - except that means another technology to add into the cost mix which generally only appeals to the high end buyer. And as I suspected from my initial analysis, there is no simple or obvious solution for the smaller business or the ISVs. They're on their own. That's not good.
We are told that Salesforce will launch a data connector in the coming weeks but we don't know how well that will work.
Back to SAP HANA. While that solution started life as a database for high speed analytics, it has morphed into a more general solution for the whole SAP landscape. That means transactions and analytics along with custom solutions. Wave isn't in the same class but Salesforee argues it doesn't matter for its buyers. We shall see but I believe that positioning will turn out to be Salesforce undoing in a number of deals and especially in the ISV situation.
ISVs I have spoken to love Wave and with good reason. You can do a lot with Wave which was previously impossible that look superb on any device. It comes at a price. ISVs will need their margin and right now we have no idea what OEM pricing looks like. That's OK to some extent but my sense is that once customers get an inkling of what Wave can offer, then ISVs could be overwhelmed with requests. This issue needs fixing in double quick time.
Operational analytics is a different beast to your normal reporting etc. Here, users get to ask many more questions which in turn means many more templates which also means more consulting for ISVs. That's not a business they'll want to be in. For all intents and purposes, the Salesforce model outsources IT to the ISVs but wy wuld they want to be in the template building business? End users will initially struggle to figure out the kinds of template they want to build which in turn adds pressure in the sales and post-sales cycles.
To make matters a bit more convoluted, I am getting slightly mixed messages from ISVs. On the one hand the visualization and filtering capabilities are great but on the other hand, the technical limitations and/or unknowns that exist today are making it virtually impossible to work with live data.
Instead, ISVs are having to dummy up small datasets with which they can show capabilities but then are left head scratching about what to do in the real world. Perhaps I've been unfortunate to hit upon ISVs that have complex datasets. Regardless, that's a problem which is supposedly being fixed with connectors to be released 'soon.' I'll wait for field reports before going any further.
Salesforce rushed out Wave as a way of getting something out into the market for Dreamforce. The product has been thought through at the user end and it is encouraging that Marc Benioff, CEO Salesforce was unequivocal that the company worked very closely with pilot customers to get this first iteration out the door. That's very smart and welcomed.
It's the next step that bothers me. With so many things left unclear and as yet unknown, buyers and ISVs are currently in the dark as to how well they can position Salesforce Wave as a truly differentiated offering. One thing is for sure - a slick front end with lots of things to twiddle with ain't it.
If I learn anything substantially different at the Wave keynote then I'll update this story.
Disclosure: Salesforce.com and SAP are premier partners at time of writing.
Image credit: Den