Using AI/ML to predict ERP Utilization

There are not many methods on how to maximize ERP utilization. There are approaches that do a commendable effort in identifying some of the factors that influence ERP utilization.  What is lacking are predictability, repeatability and harmony across the key components of a business solution.  In my 25 years of ERP consulting experience, I have never encountered a single customer that utilized over 60% of the ERP software.  Given the lack of utilization and the money needed for cloud ERP implementations, I am convinced this is a problem that finally must be solved!   We are in the second generation of ERP software and the only advancement that we in the ERP implementation arena have made is to retract our initial recommendation of customizing ERP for greater customer value.

The purpose of this article is to provide an overview of my theory on predicting ERP utilization.  The scope of this article will focus on the cloud ERP deployment model.  I have yet to complete the rigors of the scientific method to verify my theory.  However, I would like to share my concepts with you and partner with you in growing the collective knowledge.

ERP Utilization requires all components of a business solution

First we need to revisit the concept of a business solution, the key components, their relationship, and influence on ERP utilization.

Just as technology has a part to play in ERP implementation success, business process maturity and the organization’s ability to change have a greater impact on ERP utilization.  To elaborate on this thesis I will make use of two models:

1. Capability Maturity Model Integrated (CMMI).
2. Organization Capacity for Change (OCC).

I find that Organizational Change Management (OCM) is more of a project-based effort to enable an organization to meet a specific event.  This approach works very well with an on-premise ERP solution where upgrades are measured in years.  However, in the more dynamic Cloud ERP solution model, change is more rapid.  ERP cloud updates and upgrades happen in months, not years.  What I really like about OCC is the greater focus of providing the organization with the skills and flexibility to handle known and unknown changes. 

OCC is an emerging model, thus there is limited content regarding how to assess and measure OCC for an organization. There for 4 key attributes for defining OCC levels (Saylor Academy, 2012):

1. Organizational Trust
   • It refers to how much front-line workers trust middle managers and senior executives to watch out for their interests.
2. Lateral Leadership
   • Focuses on getting things done across organizational units and functional areas of expertise. Fisher and Sharp (2004).
3. Systemic Knowledge
   • Systemic knowledge is the degree to which members of an organization understand and are focused on the overall organizational system.
4. Cultural Ambidexterity
   • Change-capable organizations balance accountability with innovation. If the organization overemphasizes accountability, innovation suffers. And if innovation is the sole focus, accountability is ignored.

Given the above definitions, I have created a level definition to support my ERP utilization prediction model.

ERP Utilization Model

With all that said, consider the following conceptual model:

I propose a multivariate linear regression relationship between business process maturity, organizational capacity model for change with potential ERP utilization. This simplified model is based on the following assumptions:

1. A set of ERP features require a certain level of organizational and business process maturity for a successful experience.  For additional information, see my article on Business Leads and Technology Supports.
2. Based upon the CMMI level and OCC for the customer, we can infer the ERP features required for maximum ERP effectiveness.
3. OCC has a direct influence on effective ERP utilization.
4. Software changes will happen more rapidly in an ERP Cloud delivery model versus a tradition ERP On-Premise model.  Therefore, OCC must become an ongoing competency (versus a one-time effort) for long-term ERP success.
5. “It is generally not fruitful to impose a very sophisticated process on an organization whose maturity is low.  The maturity of an organization not only depends on the skill sets of the individuals, but also on the chemistry of the team.” (Alexia Leon, 2012)

Based on the above model, I conclude that customer enablement must be an ongoing exercise that runs in parallel or even precedes ERP automation.

Model versus Reality

I consider myself more of a pragmatist than a theorist.  Models are great to elaborate upon concept(s) for discussion and argument.  However, conceptual models are limited in the value they provide to customers if there is no method to align reality (i.e., “as is”) with the optimal path.  Consider the following illustration:

Key Points and Observations:

1. In a majority of cases, there is a difference between potential ERP utilization and actual ERP utilization experienced by customers.
2. In order to promote repeatability, there should be a logical progression that enables customers to maximize ERP utilization (i.e., roadmap).
3. To minimize ERP Total Cost of Ownership (TCO) and eliminate cost constraints, ERP Cloud vendors should provide customers with the ability to increase ERP utilization without heavily relying on consulting services.

As we continue with the model elaboration, we find that there are regions that are not practical for a given CMMI and OCC values.  Consider the following:

Key Points and Observations:

• Circles represent a subset of possible values given that the scenario occurrence is highly improbable. For example, an organization with a CMMI level 1 maturity cannot expect to utilize 100% of the features available for a Cloud ERP service.
• The permutations of CMMI and OCC levels indicate a linear relationship for targeted ERP utilization.

What Role can Machine Learning Play?

If we can define a logical model with reliable predictive results, then we can begin the journey of providing free consultative guidance to ERP customers and prospects. Let’s start simple with a prediction formula and a structured learning test set.

Following are the assumptions that I used for building the formula and training set:

1. Business process maturity (CMMI) and organizational maturity (OCC) have a linear relationship with potential ERP utilization.
2. Certain permutations of CMMI and OCC does not reflect reality (ex. Business process cannot be at a high level of maturity and a low level of organizational maturity).
3. The minimal boundary of ERP utilization is 20% and the maximum boundary is 80%.

Next, I performed individual regression analysis for each variable separate and together in order to determine if the predicting equation should use one or both variables.

Based upon the above analysis, it appears that the prediction formula utilizing both business process maturity and organizational maturity best aligns with the training set. Next step is to calculate the coefficients for both variables.

In an act of transparency, I am sharing my data model. Feel free to review and provide feedback/correction.

So what is the application or value add this model can provide?

Summary

Even with a cloud delivery model, the implementation costs associated with ERP have not dramatically decreased.  The ratio of ERP software cost to ERP implementation cost has increased from 3:1 to 6:1.  It is only a matter of time before the ERP market forces ERP vendors to drastically reduce implementation costs while maintaining a sufficient level of customer enablement.  Given the rise and general adoption for cloud ERP services, ERP utilization is becoming more strategic competitive advantage for cloud ERP vendors.  What I see as an emerging demand from the ERP market is a reliable, repeatable method for maximizing ERP utilization.  I hope that my efforts move the discussion forward.

Join the community! 16k followers across 100 countries!