Spend Matters welcomes this guest post from Matthew Holzapfel, product marketer at Tamr.
Procurement organizations are looking to deliver value beyond traditional “spent analysis” and newer forms of analytics that can help generate the intelligence to discover this untapped value.
The intelligence lies in the correlations and implications that aren’t evident if we stick to siloed analytics run by siloed functions using siloed data. In the case of spend analysis, forensically analyzing “inside-out” data such as POs and invoices will only reveal so much. Conversely, adding “outside-in” contextual information (like commodity markets, currency markets, inventory levels, sales volumes, supply chain performance, historical data and comparable spend from other areas) adds understanding to insights. It’s one thing to know historical copper prices you paid. It’s another to understand what’s happening in copper markets, why and what you should do about it.
Data unification, a new strategy for bringing data from various sources together, is particularly suited to spend management, which needs access to ERP systems, PDFs of contracts, inventory tracking systems, sales spreadsheets and more, including those external sources. All of the data sets can be treated as a common resource, and any analysis is easily repeatable on demand, at regular intervals or in response to triggers.
The first step to unifying data is cataloging the varied and dispersed data sources. Most people in any organization only have immediate access to about 10% of the company’s data. Free data catalog tools help analysts discover, organize and understand the procurement data in the organization. The value in these tools increases dramatically as their use spreads, because analysis strengthens geometrically as sources accrue linearly.
Analysts can then evaluate fields in each source relevant to their research, to establish how the data relates to data in other sources. This data connection process is no small feat. It often requires intelligence that is just as dispersed as the data itself.
Companies are relying on a combination of machine learning and expert-driven analytic models to get this matching done. Machine learning tools can identify data sets that have strong correlation to help analysts quickly combine multiple fields. Where machine learning and the analyst’s own judgment fall short, a trusted network of experts in the data are called upon to clear up issues. This, too, can be aided by machine learning — having software identify the most promising experts for each question.
Once the data is clean and ready for analysis, companies might identify when certain suppliers are most likely to increase prices when certain commodity prices change. They might analyze key drivers of non-compliant spend — or any spend for that matter. They could look for changes in terms, like delivery schedules or inventory requirements, that can predict supplier distress.
It’s time for procurement’s second act — not only in following how much money is flowing where, but also why it’s flowing, what’s driving it, how to predict it and most of all, how to improve it. Data unification is here to help.