What is spend analysis?
- 14 Jan 2020
- Spend Management
The concept of spend analysis has been around for twenty years or so. It started with the classical “spend cube” in Excel that many consulting firms used as a foundation for crafting the cost out programs they sold to CFOs and CPOs. Then it developed further with more advanced spend analytics tools and technology.
Ardent Partners’ research shows that with a superior level of spend visibility, the average Best-in-Class procurement organization reports higher levels of annual savings (6.3% vs. 6.1%) and 47% better contract compliance rates than the other groups in the market.
Many names for the same concept
Throughout the years, the same basic concept has been re-packaged, re-named, and re-marketed following trends around “big data,” “analytics,” and “digital.”
In short: Spend analytics is the process of turning invoice data to meaningful insights that lead to savings. The process can be described in five main steps that are the same regardless if you use powerful technology or a spreadsheet.
Create a spend category tree. The category tree spans across geographies, cost centers, functions, and any organizational belongings or responsibilities. Within indirect spend, the purchasing categories can be created from standards such as for example, UNSPSC. The scope of direct material varies considerably more between different industries, and thus a more tailored approach towards defining the category tree may be needed.
Identify, extract, and import data. Extract spend data from, for example, the accounts payable ledger in the ERP system or directly from the invoice management system. The data should at least include every invoice or invoice row with associated information about the supplier, date, total, currency, account, and cost center. However, all information is good information when AP professionals want to draw insights from the data. With a best in class automated accounts payable process that captures all the information on the invoice, down to the line item level, the effort in this and the two following steps in the process is considerably lower.
Data cleansing. The third step involves cleaning, correcting, and normalizing data. For example, managing errors in the supplier list and grouping duplicates, or misspelled item or supplier names.
Categorize purchases to the category tree. Once all the data is in the same format, the fourth step is to categorize the AP data. Start the categorization on the account level, i.e., certain GL accounts can be mapped with their entire spend into certain categories. Second, analyze the suppliers. Since the category tree is constructed based on how the supply market is organized, the majority of the suppliers will fit into one or a few categories. This means that the vast majority of categorization can be done by defining which category each supplier should belong to. Be aware of the cases where the supplier needs to be divided between more categories.
Having spend visibility and using spend visibility to achieve an advantage are two different things. The value of spend analysis is not the data itself, but in the ways that the data is leveraged to take decisions that drive greater value and improve performance. Thus, the fifth and final step is to analyze, conclude, and implement actions.
Spend analysis re-imagined for the future
In the last few years, the rapid increase of automation in the invoice management process and technology, such as machine learning, has revolutionized how companies can capture detailed spend data even on non-PO invoices.
Gartner predicts that “By 2022, 50% of all legacy spend analysis software will be retired; replaced by artificial intelligence (AI)-powered, cloud-based solutions.” and “by 2020, natural-language generation and AI will be standard features of 90% of modern BI platforms.”