How to avoid common mistakes in PIM projects
Many years ago, we came across an article by Michael Mezger and Marc Hölzle from Parsionate. In their text, they shared experiences and provided valuable tips on how to avoid common mistakes and complete a product information management project on time and within budget. The tips and advice in the article proved invaluable in many projects we had the opportunity to undertake later. This article became a true treasure trove of practical advice for us, so we would like to present some of them that we find particularly relevant.
Designing a system to meet all requirements for the next 25 years, preferably all in the first phase of the PIM project
Suggestion: Start by detailing the initial phase while anticipating future steps. Consider the full spectrum of services that PIM should offer in the near future. Then take a step back and begin by implementing only the basic functionalities of the software in the initial stage, proceeding in small steps thereafter. For instance, you could start with online integration, then move to a small printed project before attempting a large, comprehensive catalog. Avoid implementing a grand project that addresses all requirements for the next 25 years all at once. Otherwise, the specification and design phase might take five years, only to discover that the goals differ completely from those initially defined.
Thinking that everything existing systems cannot do must be possible with the new PIM system: the myth of the SUPERTOOL
Introducing a centralized, multifunctional system always carries the risk of raising expectations that all problems will be solved, including those for which no one was previously responsible.
Examples of special requests:
- Storing in the PIM system everything that doesn’t fit elsewhere
- Generating reports on transactional operations due to the absence of a BI system or data warehouse
- Calculating prices when, for example, the ERP system cannot handle it
Managing master data, consolidating data, and exporting it to various channels are the core features and strengths of PIM software. Striving to add too many additional functions unnecessarily complicates the system, increasing costs. Moreover, users may still not be entirely satisfied, as their expectations will likely never be fully met.
Designing an overly universal and complex data model to handle every challenge
Companies tend to create a data model capable of addressing any unknown situation, regardless of its purpose. Instead of defining specific entities, only general objects with abstract information and relationships are modeled. All information from all systems, whether external or internal, is included in the data model. Every piece of information, whether it will be used or not, is included in the data model.
Advice: Focus on actual requirements when designing the system. Otherwise, the system will become unnecessarily complex, increasing the burden of maintenance, which may become unmanageable.
Underestimating the complexity and time required for data migration when planning a project
Companies often have multiple sources from which data must be transferred but generally lack full knowledge of where their core information is stored. All these sources must be identified. The same or similar information is often stored in different systems. It is crucial to determine which data should be sourced from where or whether combining information from various sources would be more appropriate. Often, data quality in the system is overestimated, while the effort required for analysis and data cleansing is underestimated. Data migration and cleansing are usually time-consuming and error-prone. On the other hand, the time available for data migration within a given deadline is often limited. Collecting missing information and manually correcting it takes considerable time.
Suggestion: Identify data sources and assess their quality at the early stages of the project. Allocate sufficient time for data migration in the schedule and ensure this process is tested well in advance.
Customizing everything possible: Solving every exception through software. After all, standardization is only for beginners.
Another common mistake when implementing a centralized system like PIM is trying to replicate current processes literally or “improve” them in the new system. Often, there is also an attempt to automate every single exception. However, standard software delivers significant benefits only when the standard is applied as widely as possible. Customizations or extensions should only be made when there is no sensible alternative.
Recommendations:
- Start with the standard solution. Be open to rethinking and modifying current processes.
- Consider whether exceptions are truly necessary. Offer users new methods to achieve their goals instead of immediately customizing the software.
- Always check whether built-in standard features in the PIM system can solve the issue before making changes to the PIM system.
- Delegate individual tasks to other subordinate systems, for instance, for data collection or storage, to avoid modifications to the central system.
- Remember that customizing the standard solution can pose risks during subsequent PIM system updates and often involves significant costs.