When Infor announced it would be using Amazon's Redshift as the business analytics platform, I questioned whether this was appropriate given the then recent Snowden revelations. My concerns remain although now I would expand them given the role of Amazon Web Services at the center of Infor's cloud play.
Before doing so, I'd remind readers of Derek's analysis, where he points to exchanges in the media Q&A:
Someone noted that Infor’s strategy is unusual compared to the rest of the cloud market, where most players have focused on building out their own infrastructure. Would customers feel more comfortable with an Infor-owned data centre? Should Infor have considered making some infrastructure investments? Phillips dismissed these claims and said:
"Coming later to the cloud actually helped us. I’ve seen these cloud-owned data centres and they aren’t run that well because it’s hard to do everything yourself. They can’t come close to what AWS can do.
There’s also reason that the number of cloud companies that have been out there for a decade still don’t make money, they haven’t figured out how to do it yet. We just think the economics work out better."
Frank Scavo has his own take:
At least with newer SaaS providers, I'm not sure I agree that they are out building their own data centers #Inforum14
— Frank Scavo (@fscavo) September 16, 2014
That's right, but then most of those newer SaaS providers are not building end to end process based ERP solutions. All the solutions I have seen that run in AWS are either add-on or non-critical applications. Infor is different. Infor is saying that if you are in one of our verticals then you will probably be a candidate for our solutions and, if you want the cloud option then you'll be running on AWS. That may well work for small and some mid-sized companies but I find it hard to believe large companies will tolerate that approach. Here's why.
The deeper you go into the software functionality stack, the more dependent you become on the software provider. That is usually a good thing because the provider must keep on their toes to continue delivering innovation for your needs. If you are the vendor then you get to build a highly defensible moat around your intellectual property. Everyone wins. However, if you have regulatory constraints, then how you deploy becomes of greater importance.
Does Amazon care enough?
The difficulty with Amazon is that, as we have discovered, they don't care too much about your regulatory issues or privacy. They will, when required to do so, hand over data to government bodies and without your consent. That may well be a minor issue in the context of 99.9% of small and some medium sized businesses, but as we already know, governments are quick to play fast and loose with data that comes into their hands. We also know that government has a disproportionate ability to screw up the assumptions they make from the data they see.
Reviewing Amazon is not therefore a simple hygiene factor in decision making, but one of understanding and assessing the balance of probabilities that your data may end up in the hands of a government not exactly given to knowing WTF it is doing. That risk may be small, and in the past, Infor has said that it is the first line of defence for its customers. But unless it explains that in more detail AND backs it with sensible contract language then those are just words.
Phillips' argument that competitors haven't figured out data center economics is a bit wild. If you review the financials of any large SaaS vendor, the cost of data center operations is a small percentage of operational cost. The big costs come in fueling growth and it is those that drop the vendors into loss or break even. Last year, when NetSuite was considering a foray into Europe, CEO Zach Nelson pointed out that the costs of spinning up a data center are almost trivial.
That's not to say that Phillips is making the wrong argument. He may well know something about Amazon the rest of us don't but there remains another factor, which, ironically, plays back to Phillips own argument. Amazon has been trading 20 years and to date has barely shown a profit for the amount of capital it has raised. Instead, it leverages excess cashflow to continue building out facilities while trying to keep as close to break even as possible. That is a defensible strategy provided the stock price remains on an upwards trajectory. If that changes then things become difficult very quickly.
Without wishing to get into the various economic arguments around Amazon, (that's for another time) I find it difficult to believe that Phillips does not have a Plan B if Amazon's lights are suddenly dimmed. To my knowledge, no-one has asked that question yet with SaaS revenue declared at what looks like a conservative $100 million per annum run rate and a 'cloud-first' approach to future sales, then it is only fair to ask how Infor will maintain control over its cloud infrastructure.
Speaking of control - most commentators believe Amazon will end up dominating the cloud infrastructure market. Google and Microsoft might have something to say about that but what happens if Amazon decides to unilaterally increase prices? What does that do to Infor's model? At the very least, it begs another question. To what extent has Infor locked in cost for the benefit of its customers? That's another handy question on the RFP to which I haven't yet seen an answer.
I have a great deal of personal admiration for Charles Phillips and his small team of executives. Their decisions are often purposeful, logical and clear. They suggest a roadmap that's easy to understand and is well articulated. In short, they demonstrate a capacity to build trust that is unusual.
If you accept my arguments, then using Amazon represents a leap of faith. I am happy to be proven wrong but I still want comfort factors around my customers. That in turn means I want to see and hear language that provides customers with continuity assurance plus re-assurance that my corporate data is managed in my best interests.
Disclosure: Infor is a premier partner at time of writing