How do you deal with data silos and fragmentation in your spend analysis?

To address data silos and fragmentation in spend analysis, I consolidate data from all sources into a single, centralized system. I use advanced data integration tools to harmonize and standardize the information, making it easier to analyze and derive actionable insights.

Para enfrentar silos de dados na análise de gastos, comece identificando todas as fontes de dados relevantes, como sistemas ERP, software de compras, contas a pagar, extratos de cartão e contratos. Mapeie também como os dados fluem entre essas fontes, considerando a frequência e métodos de extração. Esse mapeamento ajuda a entender a complexidade dos dados e identificar possíveis lacunas ou sobreposições.

Single Source of Thruth In the era of digital transformation, having a reliable SSOT is no longer optional. It’s the compass that guides enterprises toward data-driven success, agility, and resilience. By embracing the SSOT philosophy, modern companies can navigate complexity, make better decisions, and thrive in an ever-evolving business landscape. The intention to introduce an SSOT is one thing, but the implementation is another. There are often unexpected hurdles, such as: - Existing systems may not seamlessly integrate with each other or follow standardized procedures. - Delays in data availability and inconsistent data quality - Organizations with multiple systems resulting from M&As face data integration complexities

It is a best practice to have a single source of truth for a company, and many times that system of record is a compilation of other systems. That way, people looking for information only have to look in one place and any conclusions they draw or decisions they make will be based on a complete understanding of all relevant data - regardless of what system or part of the company they are connected to.

Cleansing and standardizing data is an important step in making data more actionable and trustworthy. The classic example of this is a system that lists "IBM" "I.B.M." and International Business Machines as three separate entities. Although it is a large task harmonizing all of the data created in the course of business, this is a perfect place to leverage digital platforms and even AI, only having humans get involved in the case of exceptions.

Shit in Shit out. It is not without reason that some BI tools have an ETL tool integrated with which you can first harmonize different data sources. The quality of your data largely determines the end result. Who hasn't experienced it when you have different sources and one uses a dot and the other a comma as the thousands separator. A small detail that has a huge impact.

While standard schema, such as UNSPSC, are common, many of them were not created for corporate purposes. Fortunately, modern spend analytics platforms make it easy to have multiple taxonomies. One of these should absolutely be a custom schema created by the company for its own use. This may sound like a lot of work, but it does not have to have more than 2 levels of categorization in order to be an effective - and familiar - way of categorizing and reporting on data.

One of the traps that procurement often falls into when doing analysis is falling back on Excel... while this is a convenient and comfortable tool to use, doing so creates risk and actually takes more work in the long run. By far, the preferable option is to conduct analysis in the platform where the data is stored. This way, everyone can access it and it can be updated easily as new data becomes available.

One that I would never had of used in the past is to also work with purchasing groups who help to manage spend on certain categories where you either do not have the resources or the time to address. Allow them to help you get that spend understood and under control moving forward with moving it into the correct spend silos. You can gain short term savings and understanding of fragmented or hidden spend.