Nov 17, 2018
I was listening to Gus Gonzales' recent podcast on Ten Features that should be removed from Dynamics 365. As usual, Gus makes some great points, but when he was discussing limits, I think he left out a part of the story, that leads to a larger question.
We were recently engaged in a P2P scenario, where a partner had quoted a fixed price to their customer, to perform some work that they needed our help with. We were engaged by the partner on a hourly basis to perform this work. The project turned out to be significantly more complex than the partner had imagined when they quoted the fixed price. Sure enough, our hourly cost to the partner ultimately exceeded their fixed cost from their customer. This was a cost mismatch that was unfortunate for that partner.
While Microsoft aspires to the "One Microsoft" story, they fact is that there is not "one" P&L. Each business group has it's own P&L, which is necessary for Wall Street and many other reasons. When one Business Group uses services from another Business Group, they are not free... they get a bill, just like we do.
With version 9 of Dynamics 365, there was a lot of hoopla about the legacy CRM database moving to Azure. It was going to be so much better, more scalable, faster, etc. And it was indeed all of those things. But there was also one new wrinkle. The Business Applications group was now going to receive a bill from the Azure Business Group for all of those spinning servers and storage services. Azure is sold on consumption, but Dynamics is offered on fixed licensing. Sounds familiar to my P2P story.
I agree with Gus that having a 50 record aggregation limit can occasionally be an issue. While Gus feels his customer might be willing to wait an hour or two for the results of a million record aggregation to be complied, that is going to spin the Azure meters like crazy. That customer with a need for a million record aggregation, is paying Microsoft the same amount per user, as customers who never exceed 50K record aggregations. But Microsoft's COGS could potentially eclipse the revenue that customer is paying. This is not a good spot, it is a mismatch.
One way to fix the mismatch, would be for Microsoft to change the way they offer Business Applications from Licensing to Consumption. In a consumption based model, there would be no need for any limits, on anything. The customer could spin the servers, and save as much data as they wanted, as long as they were willing to pay for it.
Some of you may recall my battles with the Dynamics 365 for Marketing initial Licensing scheme. What I was really lobbying for was a consumption model. We ultimately landed on a workable compromise, which is kind of a quasi-consumption. I recall, one of the obstacles to going to a straight consumption model was "That is not how any of our competitors offer this". I was of course thinking that was exactly why we should do it.
Let's imagine that you are renting an apartment. You have a fixed monthly rent amount (License), but your electric bill varies (Consumption). While I cannot control my rent amount, I know I can control my electric bill, turn off some lights, adjust thermostat etc. When I go out of town for a month, my electric bill drops way down, but my rent does not. Imagine if instead of paying a fixed monthly rent amount, there were thermal sensors in the apartment, and was only charged rent based on the number of hours someone was in it. I would not pay any rent for my two week family vacation. Would people buy that? BTW, Microsoft IoT would be a great choice for any landlords considering this approach.
This is not unique to Microsoft Business Applications. Netflix charges a fixed fee to customers, but pays Amazon on consumption. When everybody started sharing their Netflix login with friends and family, Netflix was getting hammered, so they introduced limits. The opposite is also true. The licensing prices for Microsoft Business Applications are based on some formulas of typical usage. Some users are above the range, but those below, are in effect, subsidizing those above. Same as how skinny people offset the fat people at a Chinese buffet. The risk for the buffet owner is the local chapter of the "Society of Obese People", picking their place to meet. The recent exploration of changing the way Microsoft charges for instances and storage was an attempt to level this out. Microsoft temporarily stepped back from the consumption based storage model, but what if they came back to it... with everything?
Imagine that everything under the Microsoft Business Applications umbrella shifted from licensing to consumption. Use whatever you want, as much as you want, with no limits. No figuring out what licenses users need, instead maybe you stick meters on roles. It would not have to be tied directly to Azure consumption, but some sort of metric(s) that go up and down based on usage. Logins, record views, number of records in an aggregation, etc. If a customer closed the plant for a month for the holidays, their cost would be near zero. When they ramped up activity for back to school, their cost would be higher. A direct relationship of use vs. cost. No one subsidizing anyone else. Could you sell that?
Yup, I hear this argument. "Customers will always want to know exactly how much it will cost!". I think that is an assumption many are making on behalf of all customers, that has proven not to be true for other things. One of the fastest growing services on the planet today is Microsoft Azure, with no fixed cost. Azure sits right next to another service, from Amazon, that also has no fixed cost. Both of those tools have "cost estimators", which is as good an idea as you're going to get. It appears obvious to me, that customers are more interested in paying for what they use... or at least the "skinny" customers.
If the cloud is, as Gavriella Schuster put it to me recently, the "Fourth Industrial Revolution", then consumption models will be the fuel. By the time your high school age children, have children of their own, this will be the norm.