Recently, diginomica’s Jon Reed and I spoke with Jon Allen, CIO/CISO of Baylor University. I had interviewed Allen previously in 2019 when Baylor was starting their Oracle Fusion Cloud HCM implementation. That solution went live in June 2020 in spite of the pandemic challenges most organizations were facing.
This week’s conversation was instructive as it provided an opportunity to find out what did and didn’t work. It also pointed out some significant planning and headcount matters that almost all cloud application software customers should heed. This piece will focus on the latter points.
One bit of context first. Baylor has implemented a large suite of HCM applications. Why? According to Allen:
We came from a world where I think on the HR side alone, we had 18 separate systems to execute our HR operation, everything from employee recruitment, to onboarding to training, it was a fragmented mess.
Implementing all of those modules would challenge most any large entity, but it turns out that implementing cloud ERP products may have some long-tail activities that warrant additional investigation and staffing.
The implementation focus today
ERP vendors - and many of their partners - are often focused on fast implementations. A fast install may add value in some situations but it may not deliver much else if it doesn’t address process improvement opportunities, the use of advanced technologies (eg: Big Data, ML/AI, bots, etc.), foreseeable business and technology changes, new workforce requirements and more.
Fast installations, particularly those where data from the old system is merely mapped into the new system, may simply replicate old functionality, processes, reporting, etc. in a new piece of software. Given how long many firms keep their old ERP solutions (some marriages don’t last as long as an ERP installation), the time to improve things is during the selection and implementation design – not after it’s too late.
But, our conversation with Allen went well beyond these points. We got a masterclass in the staffing and issues that must be considered in a post-cloud solution implementation world. The core advice he offers is to recommend that ERP buyers develop a different staffing model than what they had when they had on-premises applications. Read on to see the new model.
The team you really need
The first concept Allen identified is that cloud/SaaS software should be thought of as part of a continuous improvement process and you’ll need people focused on this. He told us:
I think that really comes to light as you enter the post go-live world. You soon see that you just don't have the static world that you lived in before. It’s a very dynamic, continuous improvement culture that your organization will need to adjust to. As I've talked to campuses across the country this is one of the key challenges that I have seen. People have treated these projects like projects: You go in, map your business processes, map the processes to a system, get to the end, and everybody thinks they are done. And then they realize this new world is nothing like the one before where these processes stay static for the next five to 10 years. That's not the world we live in anymore.
We now operate in the mode of continuous improvement driven by evolving business processes, undergirded by modern technology that is constantly adapting to business needs. I think that's the piece that all of the sales teams haven't necessarily done a great job of highlighting. I think there are better narratives coming from vendors on that conversation now than there were before because of the lessons we are learning.
OK, so the business world and your firm are constantly changing. Do we really need a continuous process improvement group? You likely will as both the business environment AND the technology are both rapidly changing.
Minimally, what SaaS ERP buyers should be doing is:
- creating a different way to evaluate new releases
- identifying the potential value they can have on the organization
- understanding what will be required to successfully implement them
- developing a comprehensive upgrade plan that extends over several time periods
As to why this new capability and why now, we need to remember how things changed when on-premises applications were displaced by cloud/SaaS apps. Software buyers who were used to seeing updates every 18 months are now getting batches of security, performance and functional improvements every quarter now. The volume of changes has grown significantly and customers need a mechanism and staffing to deal with this.
According to Allen:
One of the things that happened for us post go-live is that we weren't able to manage the system using our existing support infrastructure. As a result, we retained our implementation partner to help us navigate quarterly updates. With every update there are new features available, but maybe the existing feature isn't operationally deprecated yet. So it could be two or three quarters before the existing way of doing business is deprecated.
The challenge is if you aren't pipelining this right within your organization - understanding that it isn't just about this release but what do we need to be doing over the next three to four releases – then you can really get caught by surprise. Within three quarterly releases your organization will need to quickly transition to new ways of doing things, training users on a new responsive interface, adopting new user-facing functionality. The speed at which all of this happens is very different from the traditional way things were being done. That's a big piece of the puzzle. In my experience, universities haven’t had areas of the organization thinking about continuous improvement pipelining.
I've told Oracle, I've told consultants, I told everybody, that it never takes fewer resources to run these cloud systems. For those thinking about this transition that have been told moving to a cloud ERP will reduce the load on IT, that’s a lie. It's flat out a lie.
Process changes should be part of every implementation and will continue to be revisited as new product releases occur. One example Allen gave highlights the importance:
You know, onboarding employees used to take 26 steps. It was ridiculous. It was hard to track. It was fragmented, that has reduced dramatically.
Allen also highlighted one group of software users that create special change, training and follow-up challenges: infrequent users. He states:
Some people may do some kind of data processing once a month and do only a handful of transactions. If you do something rarely, you're not going to remember how to do it. There are going to be inaccuracies and other issues that arise. We're starting to peel that back. Let's actually get to the data and understand why we are seeing challenges. Is it a business process? Is it a training issue? Is it lack of specialization? Is this an area where we can have more specialized staff and a more centralized way of executing this work to free up resources out in schools, colleges, administrative areas to do their core business? These are some of the questions we are asking now. The answers to these require a culture of continuous improvement. Organizations that aren’t asking questions like these are going to get caught by surprise.
What Allen is saying is that cloud ERP selection and implementation teams must plan for the headcount, processes and business changes that will be factors in the post-implementation world. While the quotes above point to a couple of post-implementation needs, the firm will likely need skills in other areas, too. In short, every firm should have people for:
- Continuous Process Improvement
- Release Management
- Process Reengineering/Process Automation
- Integration Management
- Data Management
- Identity Management/Security
- AI/ML/Algorithm Tuning/Management
And all of these are in addition to the functional headcount you need to actually process transactions, deal with exceptions, etc.
Allen also discussed how change management will be a discipline that is needed throughout the cloud application usage cycle and not just during the implementation phase. He stated:
I think that's a piece that people are starting to understand. As I’ve collaborated with other institutions who have had successful and challenging implementations a recurring theme is the commitment of leadership to cultural change. Technical challenges are almost never the reason implementation projects struggle. The biggest challenges have to do with the cultural changes that must occur in the process of making business improvements.
I’d agree and add that the number and potential software changes that are possible will put a strain on firms trying to decide which capabilities to accept and how to successfully roll these out to potentially thousands of users. This is not an insignificant matter.
To help with this, one tool Allen found helpful was Oracle’s Guided Learning:
Oracle Guided Learning was something that we hadn't even considered during our project. But when COVID-19 hit, we ended up licensing that tool. It was extremely helpful in guiding users when we couldn’t do in-person training and where online training was a challenge as we adapted to everything being online.
Oracle Guided Learning allowed us not only to help people point and click their way through new procedures, but to customize that information so that it aligned with what was happening in our business processes and institutional needs. While there are a growing number of players in this space, the fact that Oracle has a tool that adapts with each quarterly update has been a big win for us, especially as our operations shifted to remote for a period of time.
Allen also commented on the criticality of integration and why great skills, dedicated staff and tools are needed here. He stated:
One of the things that was surprising to me and most of my peers is the importance of integration teams. I don't know how many universities had an integration team or an integration platform 10 years ago, but if you don't have that today and shift to a cloud ERP environment, I can guarantee you will fail.
Integration is the core resource that makes all of these systems work. I always joke that our integration team was a combination of SQL Connect and dumping flat files all over the place. From a CISO perspective that was a horrible approach that ultimately didn't work. So integrations are really, really key in this new world. I'm not sure that everybody has grasped that as much as they needed to going into these projects.
The other thing people don't realize is that in the on-premises world you built an integration and then forgot about it. Maybe once every couple of years there was a major update that required you to check on it. In the cloud ERP world you have to test your integrations every quarter, so organizations must have a robust process in place to do data validation very quickly or you will not be successful. Vendors are not intentionally trying to change things. You make assumptions when you build an integration. What if those assumptions change? What if the system changes something that shifts those assumptions? You've got to be on top of this.
Touching on data and identity management, Allen stated:
I currently oversee an environment where I have Oracle Cloud HCM for HR, Ellucian for students, and Blackbaud CRM for alumni. So when I think about business processes and operations that involve people, they could exist in all three of these systems - how do we manage identity? And how are we thinking about different data elements? The cloud has emphasized the need for strong data management and data dictionaries in a way that didn’t exist in the on-prem world. Some of this is motivated by the shift to specialized systems, but it is also driven by the democratized access to data in newer environments. So, there are three core areas: identity management, data management and integrations. If you can't check these three boxes and feel confident that you're in good shape, you probably shouldn’t begin a transition to a cloud ERP.
So, how is all of this coming together at Baylor with their Oracle Cloud implementation – called Ignite? Allen says they’ve got a dedicated executive and team focused on these matters:
The Ignite support organization at Baylor encompasses change management, continuous improvement, and operations, primarily involving people outside of IT. The support structure emphases functional processes – the people. Our IT organization is still very involved in security, workflow, and environment management. But this is all new! We are currently adding full-time positions to our Ignite support team. These Product Support Analysts for HCM and ERP are experts on workflows, processes and configurations, and they partner with their functional areas of the university. We are also developing resources and staffing around change management. How do we make sure that releases are properly trained out? How do we make sure that communication is good? Having these resources in place is even more important in cloud ERP environments.
Baylor is also using Oracle’s new Analytics Warehouse (note: see this article re: how Oracle NetSuite will be deploying this tool with the Oracle Autonomous DB). Jon Reed asked if Baylor was rethinking its KPIs as they re-imagine processes and rollout new HCM/ERP capabilities. Allen stated:
Yes, the first thing is taking the delivered KPIs and making sure that the assumptions they are built on align with your data and your processes. That's step one. We implemented Analytics Warehouse this summer and are in what I see as the second phase of this step, which is making it align to our data and assumptions. Then the next step is, what KPIs are missing? What are the things that aren't a deliberate KPI? The challenges are just like everything in this world: every quarter Oracle is releasing new data and new KPIs into that platform. So that's another place where you're going to be constantly re-evaluating. Is the KPI that we built the one we want to keep going with or is there a delivered one that we wouldn't have to maintain? So that will be another area that calls for continuous improvement, monitoring and evaluation. How is the product meeting our needs? What has change? What do we need to adapt?
Reed followed up with this question:
Are there any that that are really standing out for you that you'd like to check regularly?
I think there are 265 delivered KPIs today. Relative to procurement, about 80% of those out of the box, were helpful and produce good data for us. If I had to pay a consultant to go and build that number of KPIs in the Analytics Warehouse that would be a seven-figure project. And that's not even trying to maintain it!
I think that's the value proposition. People will say, ‘Well, not all of these work for me.’ But look at what you're getting and how easy the tools are to use to create your own. This isn't building pipelines. This isn't building data schemas. Everything is fully documented. For me, that's a huge differentiator and something that is unique in this space.
I think Allen’s spot-on in insisting that there are roles that must be continuously supported post-implementation as things will definitely keep changing. I also believe that too few software buyers are thinking along these lines. And, that folks, will likely lead to disappointed users, underwhelmed executives, and possibly, CIO turnover.
Smarter business executives are already realizing this and looking for new insights, new IT and function staffing plans, and, of course, budget for same. The real metric we all should be focused on isn’t necessarily the speed with which a system is implemented but how well it continues to create new, additive value throughout the life of the cloud subscription. Furthermore, I’d suggest IT professionals might want to revisit project management training and certification education as it the content may still be tied to a world decidedly more on-premises, one-and-done, world and not the constantly mutating cloud apps world of today.
Before you buy and implement your next piece of cloud application software, take some time with your IT and functional team members to really rethink how the company will get value from the implementation AND post-implementation environment. If the answers are decidedly focused on just the initial implementation, it may be time to rethink what it is you’re really going to accomplish (or not!).