Opinion Microsoft comes to their senses with Dynamics Operations On Premise

Opinion Microsoft comes to their senses with Dynamics Operations On Premise

Imagine the scene at AXUG last year. The place is flocked with dedicated Dynamics enthusiasts and the hottest question of all is about when will Dynamics Operations be offered on premise. Now, imagine that a Microsoft Representative gets up and explains that there will be no on premise version and to get used to it. In the future, there will be some on premise components (manufacturing and retail) but on premise is dead. Customers begin shaking their heads in disapproval as the Microsoft speaker, reading this, becomes even more forceful with angering his client base. The end message by the speaker, as perceived by multiple people in the audience, was “to take it or leave it”. The fallout was instant, and a number of clients as well as partners begin to seriously consider the possibility of leaving Microsoft. In fact, a recent poll by Computer Futures showed that 40% of Microsoft Dynamics Operations customers stated they were going to leave if an on premise product wasn’t offered within the next 3 years.  But all is well with the recent announcement that an On-Premise version is coming in June.   I’m pretty sure that Microsoft wishes they could take back what they said at that presentation.


Now, imagine that Oracle and several other ERP companies told their clientele the exact same thing, so this wasn’t just a Microsoft thing. There was one glaring exception in all of this – SAP. In their case, they wouldn’t dare risk offending their customer base who has substantial investments in on premise hardware and existing systems.


SAP was the one exception. They offered an on premise version and a cloud version. They also began reaching out to a number of angry customers in other product lines who felt like their software ERP vendor had abandoned them. This essentially led to a very fast reversal of position amongst the cloud only offerings to quickly add on premise versions.


There are two main reasons for this besides most everyone else was doing it. First, Microsoft and Oracle had observed how a number of companies like Salesforce and NetSuite had seemingly grown from nowhere. The key reason for this was ease of entry. Within minutes, a company can be up and running with the cloud. The lack of an upfront cost really contributed to growth and on premise has nothing that  assists the sales process like the ease of entry of the cloud.

Second, it’s the normal thing with investor expectations. Tech companies like Microsoft have it tough. It’s not like being in some of the older industries where the main expectation is stable income and good P/E ratios. Tech companies are expected to grow and grow fast. When you’ve been around as long as Microsoft, that isn’t always easy to do. A lot of products really are hitting a senior lifecycle stage. Thus, you have to introduce new lines – even if the line cannibalizes on the profits of other lines. What a lot of industry research analysts have really been wanting to see is for Microsoft to become like an IBM and really build up the services profits as that is deemed to be more stable than the quick lifecycle of software. Note, the same analysts are also the first ones to blame a company for alienating a customer base when they get to aggressive with following a recommended trend – aka “losing Core Values” as Steve Jobs Warned against when he rescued Apple(who famously was on the verge of collapse from chasing too many trends at one point).

Tickets Please for the Fight of the Century: Ease of Entry versus Software as a Service

Software As A Service(SAAS), like infamous political terms like “progress“, is one of those terms that differs widely depending on who uses it. At a general level, let’s just say that it means a company taking over more of the duties that traditionally would require existing staff. A new company buying Dynamics Operations wouldn’t need a system Administrator or a DBA as Microsoft would handle all of that for you. Yes, the companies would pay more for the software but they would save money based on not having to pay for the employee DBA, employee Network Infrastructure consultant, consultant Performance Tuning Specialist, Day to Day System Administrator. Also, the software should make things far easier and efficient. Machines are just setup instantly. People in the business never have to worry about getting hacked because the software and software provider have their backs. Performance problems and traditional ERP tuning just go away under this model. It’s so nice and so simple – all problems for the last 30 years will just go away because we used a new “buzzword”.

So, everyone wins on the profits.

Oooops, History has been forgotten… Again

With ERP systems, there is always a big risk on first time implementation or upgrades between far off versions– there is just something about taking that old custom written application that has been modified for the last 20 years and adopted all so perfectly to suit the needs of that one company. There is no guarantee that the new system will have the same success or meet the needs like the old application. So, somewhere in the late 70’s and early 80’s, ERP software vendors realized that the real money was in producing the software product and limiting the liability. For example, if an implementation fails, it isn’t uncommon for a partner to have to provide a refund. Over the years, there have been a number of partners who grew too fast, started to decline on their quality, and went broke from too many failed implementations. It happens. But the one thing that remained constant is that the software vendor still made money – unlike with service based partners, Microsoft wasn’t on the hook for failures. This is why companies like IBM tailored their strategy away from “ERP Upgrades” and more towards becoming an “Integrator” as a principal software as a Service strategy.

What gets partners and companies in trouble is having to carry the high expenses associated with having a lot of highly specialized personnel for a project. It’s not unheard of for a partner to incur over a million dollars in costs for overhead during an implementation, and many clients will negotiate a refund clause in the event of a failure. In that case, the partner carry’s significant liability and risk. That’s just the nature of the service industry.

What Was Good

The new software is no less than an absolute marvel in software engineering. Microsoft did a very impressive job. An extensible integration framework, testing framework, and development framework in general have been introduced which will pay tremendous dividends in the future. It’s much easier to now integrate Dynamics Operations with other devices and that is powerful. I love the ease with using these new retail and shipping devices. Furthermore, some of the BI tools could really be difference makers over traditionally more expensive alternatives. And for the small company using a first time ERP system, there are compelling reasons to start with the cloud – lack of investment in hardware, rapid startup, subscriptions that grow well with an organization. The ability to have a trial without any sales pressure is also critical for trying out the product. Furthermore, we all know Microsoft – a company that has never been shy about releasing something not ready and working out the kinks while customers are live. They’ve put a lot of good things in place. We are going to see a lot of benefits for the next 10 years based on these changes. And not to mention, the benefits of the marketplace. Now, customers can simply look online and see all of the ISV’s that are available. Furthermore, customers can rapidly deploy many of the ISV’s to a test environment. This is a great competitive advantage when it comes to taking control of an implementation and choosing options.

What Wasn’t as Good, Probably for Legal Reasons

So, Software as a Service must still perform according to an expected level of Service and this has to be defined very carefully. The most controversial way of doing this is the infamous, current Microsoft 8 hour response time. Code moves that previously could be done with a 2 hour outage in the previous on premise version, now take 8 hours to give the Microsoft representative sufficient time to restore the system in the event of a failure,  covering all legal basis and contracts. Taking 8 hours of downtime where we shut down an ERP mission critical system defeats the entire purpose of the new enhancements to the development framework, which clearly were made with the principles of continuous improvement and highly agile implementations in mind.  A number of hardware metrics are now hidden and considered to be strictly in the view of premier support. If a customer has a problem like a lock which they previously could just have an employee come on and fix within a few minutes, they now need to wait 8 hours. Advanced Performance Tuning, like the type that I typically do, doesn’t even exist – where people like me take advanced implementations and tune performance with competitive advantage in mind with nonstandard, undocumented approaches. It’s mostly DynaPerf and the basics, which often are not enough on the truly high volume, complex cases. Also, the luxury of having that employee who intimately knows the system and several quick fixes just disappears. Now, someone is calling a call center, waiting on the phone, and getting an 8 hour response time (not resolution which can go much longer, response). For a number of reasons, the problems really became noticeable with bigger implementations – the very target audience that everyone is chasing. For example, bigger implementations have often made extensive investments in hardware that far outperform the cloud hardware vendors. This is a small percentage of implementations but an important one. Going to the cloud means an instant performance problem when going down on hardware performance.

The danger of arguments is also there.  Now, you get into these arguments about who is responsible for what and who could get sued for what.  When an issue is truly urgent, there isn’t always time for that.

And ultimately, you understand that the real problem with this implementation of the cloud was Loss Of Control. Just imagine what would happen if all production stopped for 8 hours because of waiting for someone to resolve a lock and then having a legal contract not classifying it as downtime.  I don’t need to say anymore.  In future versions, I think that we’ll see this improve.

In Comes the Hybrid, Our Hero with the BEST of Both Worlds

For larger companies at present or upgrades from long-time Operations/AX implementations– and please date this message because this could change — hybrid seems to be the best way to go for now. There is no need to let legal battles and word interpreting stop legitimate work. The implementations that I’ve been on that achieved the most didn’t come at me with legal wording.   It didn’t matter who has messed up or who should have fixed it at one time. All that mattered is that we needed to get something accomplished and that was the focus. It’s amazing how much that works. Plus, you get the benefit of the extensible architecture and new framework – these are essential in that they actually translate to real competitive advantage.

But What About Losing BI

Stay tuned on this. At present, some Microsoft BI solutions are only offered in the cloud so they say.  Everybody was doing that also, but notice how IBM Watserv and a number of other industry leaders have just started advertising and offering their BI solutions “On-Premise”. The BI tools amongst the major big company vendors aren’t that different at present. There is a fierce competition in the BI space right now. I’m a Microsoft outsider, but I’d be very surprised if Microsoft was willing to miss out on marketshare because they refused to offer the BI platform on-premise. At any rate, we will see. Even if they do choose not to offer the financial services via the cloud, there are several other products that can offer this at present.


So, with honesty, I’d say that I really like the cloud for first time ERP implementations that are small or ones that have just a few people using AX/Operations. There are several of these in the Operations/AX space. For complicated manufacturing implementations, major companies with lots of licenses, or rapidly growing companies that are increasing their marketshare, I’d recommend the on premise hybrid or a smaller cloud solution provider (they are an exception and often assume some of the liability) when it comes out. Uptime(system being slow or unresponsive can also be downtime) and dedicated support are critical here, plus the need for faster code moves and more response times in line with dedicated support personnel. So, hopefully, we’ve gained an honest view of how to go about making a decision on whether to wait for on premise or go with a pure cloud version of Dynamics Operations.

Till the next time!