insightsoftware

The company has a consulting and implementation services team and prefers to carry out customer deployments using this team rather than generic integration partners.

The company also has technology partnerships with a company called SiteLucent for digital shelf analytics, and QlikTag for digital product passports, which are gradually being required within the EU. These contain information about the materials used within a product, recycling instructions, safety information and environmental impact.

Fig 2 - Digital Product Passport

The PIM market is a $15 billion market, growing at around 15%, and Agility is exceeding this market average, growing at around 20% annually in recent years. Its large parent company gives it the ability to consider acquisitions in addition to organic growth. It spends a healthy 12% of its revenues on research and development, and this can be seen in the latest version of the product. This has a well-crafted user interface which allows for elaborate searching and filtering of products in the system, based on the open-source Elasticsearch engine. One of the most time-consuming aspects of PIM is the production of product descriptions, which need to be quite complete for many electronic marketplaces such as Amazon. Rather than relying on these being entirely written by customers, Agility has an AI generation facility that utilises ChatGPT (or other large language models) to be used to generate product description text. Unlike some other tools, Agility has designed an interface that accesses the PIM data and from this generates a tailored, highly detailed prompt from an extensive template. In this way the users are not just expected to have a free hand with a prompt interface; the product goes a long way to providing a carefully tailored prompt to generate a description of the desired length. This draft description can then be reviewed by a human being to ensure accuracy. Of course, large language models can hallucinate, but the very detailed prompt engineering used here minimizes this as much as possible, and also allows human review of the draft content. This is a good example of the research and development work that has been done in Agility, and is one of the better-engineered examples of this that I have seen.

The bottom line

The Agility software is a well-proven and mature PIM solution that can be deployed on public cloud, private cloud, on-premises or in a hybrid configuration. It has a flexible data model and extensive built-in PIM functionality. It also has a parent company with deep pockets and has been growing faster than the market average. Mid-market customers should certainly consider Agility for their PIM needs.

Figure 1 – The data warehouse lifecycle

The key component of the Magnitude Information Engine is the Business Modeller. A partial screenshot is shown in Figure 1. The first point to note is that this has been designed for business users and there is no requirement for either a logical or conceptual model, though you can import models from ERWin or PowerDesigner if you wish. While we will not discuss this model in detail, it is worth mentioning that black lines represent requirements (a client must belong to an age group), may or may not have a credit rating, and the recursive relationship with credit rating indicates support for ragged hierarchies. Multi-value and bi-directional relationships are also supported.

The architecture of the product is illustrated in Figure 2 and a significant feature is that the Magnitude Information Engine is commonly used (around 70% of customers) in conjunction with Magnitude’s MDM solution. The two are closely integrated, with hierarchy management, workflow, master data stewardship and governance capabilities provided to ensure data is organised and validated for reporting purposes.

Figure 2 – Architecture of the Magnitude Information Engine

The way that it actually works is that transactional data such as sales figures is stored in a conventional star (or other) schema but contextual information such as product hierarchies, organisational structures and so on are stored separately. It is this separation that gives the product its ability to manage changes in the business context without affecting the data warehouse structure. Further, change management is not limited to individual data warehouses but applies also to federated environments and to the propagation of changes to data marts. This approach also results in flexibility in the way that business structures are supported. In particular, multiple structures: historic, current and planned, can be supported concurrently, which means that, to use the company’s own parlance, you can keep a true “corporate memory”.

Creating data warehouses is a time consuming and costly business, so automating as much of that process as possible, is a good idea. In addition, it has traditionally been the case that the principal reason for the failure of, or delay to, data warehousing projects has been in assuring good quality and trustworthy data. It therefore makes a lot of sense to have data warehouse automation integrated with master data management though, of course, you can still implement either of Magnitude’s relevant offerings on its own. With specific reference to data warehouse automation we wrote the following when we reviewed what was then Kalido back in 2001: “where the product is unique is on its emphasis that everything will change. You will acquire new subsidiaries, merge with someone else, add new products with new product attributes, add new sales structures, change your reporting methods and generally adapt to meet new business demands. Traditional data warehouses are not well able to cope with such change. Kalido, on the other hand, has been specifically designed to manage changes just like these.” We quote this because it remains valid today. Conventionally, when a business model changes, for example when your organisation takes over another company, or when there is internal restructuring, then the data warehouse has to be re-organised, a process that is often time consuming and complex and which can take many months. Using a tool such as Magnitude Information Engine enables changes to the business model without impacting on the data warehouse and, in practice, it enables the redeployment of a new version of the data warehouse within what may only be a matter of weeks.

The Bottom Line

There are two major issues that Magnitude data warehouse automation addresses. The first is rapid implementation, and the second is support for change. Both of these are important and should result in significant time and cost savings, as well as increased productivity and time to value.