Recently a client asked if they should keep their existing ERP software or consider replacing it. Like many companies, they had purchased various software products over the years to tackle different problems in their environment. These purchases inevitably resulted in creating silos where it is hard to keep duplicated information synchronized. Old software meant new hires had no experience in that technology (think green screen terminals) and needed extensive training. Also, reporting was simply so much work that the business didn\u2019t get the information they wanted when they needed it. The client\u2019s question was this: Do they add something like a business intelligence layer over the top of the existing systems, or is it time to consider upgrading to a modern ERP?\nJack Welch once said that \u201cIf the rate of change on the outside exceeds the rate of change on the inside, the end is near.\u201d The question faced by companies in this position is whether existing systems can scale up to accommodate planned growth, or will their limitations constrict growth? Even if current systems can scale up in the short term, what are the prospects in the medium and long term?\nGiven the rate of change in business today, there will come a point at which the existing collection of systems just can\u2019t accommodate the organization\u2019s needs. \u00a0For example, can those applications support new governmental regulations? With natural attrition, there is a growing gap between new and existing employees. Do they really want to invest in training new employees on obsolete applications?\u00a0 Therefore, the question becomes not \u201cShould we change?\u201d, but \u201cWhen should we change?\u201d\nOne way to answer this question is to seek out experienced opinion, but opinions are always biased. There is no way to avoid bias because everybody is limited to their experience. Given how comprehensive ERP software is, that experience will be either broad or deep, but almost never both. Experience is also current or old, and given the rate of change in the market, experience gets old very quickly.\nThe way to overcome bias is to measure how well existing software meets requirements and compare this with potential replacement systems. If the current system scores close to those of potential replacements, there is little pressure to change. On the other hand, if there is a large gap between current and potential new system scores, that suggests the change should be sooner rather than later.\nStart with the foundation\nThe first step in answering the \u201ckeep or replace\u201d question is to create a comprehensive list of requirements. Since ERP is being considered, this will run into thousands of requirements.\nAn inadequate analysis can lead to the wrong answer to the question. \u00a0Should the client decide on a new ERP system, any significant requirements that were not discovered during the initial analysis phase will be discovered during implementation. If too many significant requirements are missed during the analysis, it is entirely possible to select the wrong software. Even if best-fit software is selected, every \u201cnew\u201d requirement discovered that can\u2019t be met \u201cout of the box\u201d means more implementation work and associated costs. Requirements that should have been discovered during the requirements analysis phase are the primary cause of implementations that are late and go over budget.\nWhen it comes to developing requirements, start by examining the features of the existing software that are used, and rewriting them as requirements. This process is also called reverse engineering requirements from features. It is critical to use this same technique on several potential replacement systems because that is how you find requirements you don\u2019t know you need. The kind of requirements where people look at them and say \u201cHmm, never thought of that. But I can see where it would be critical to\u2026\u201d\nOnce you have a comprehensive list of requirements, you need to rate them for importance to your organization. This means capturing who wants a requirement, why they want it and how important it is to them. When you have completed this, you will have developed your requirements profile, which is the foundation for making the best decision for your particular circumstances.\nMeasure the gap\nWith the requirements profile complete, you can measure how well current and potential software meets the needs. Start by using the requirements profile to create RFIs or RFPs which are sent to potential vendors. Use the completed RFI or RFP to do the gap analysis.\nUsing a normalized scoring system where 100% means a perfect match with your requirements, you might find that your existing system scores around 85%, whereas the best new systems score around 90%. That suggests that not much would be gained by doing the project now. On the other hand, if the existing systems come in at 65% and the best new systems score around 90%, that makes a strong case for starting the project now. If a new project is on the cards, double check your analysis by estimating the project\u2019s ROI.\nRecommendation\nLet me emphasize how important it is to reverse engineer requirements from multiple products. If you don\u2019t have enough requirements created from the features of potential new products, the current system will score too high on the gap analysis. Although you will have gathered requirements based on pain points of the existing software, you need the features of new products to help you discover the requirements you don't know you need. See previous article: How to know when you have all the requirements to properly select enterprise software.\nTo gather adequate requirements is a lot of work, but to quote our client: \u201cSoftware selection is like painting a building. The real work is in the preparation, not the selection.\u201d For this client, using a gap analysis to measure how well the current and potential products meet their requirements profile leads to a rational, data-driven decision of if and when to replace their existing ERP system.