Spend Analytics Use Cases: AI & Data Science

In this post, you will learn about the high-level concepts of spend analytics in relation to procurement and how data science / machine learning & AI can be used to extract actionable insights as part of spend analytics. This will be useful for procurement professionals such as category managers, sourcing managers, and procurement analytics stakeholders looking to understand the concepts of spend analytics and how they can drive decisions based on spend analytics.

What is Spend Analytics?

Simply speaking, spend analytics is about performing systematic computational analysis to extract actionable insights from spend and savings data across different categories of spends in order to achieve desired business outcomes such as cost savings, cost avoidance, spend forecasting, spend leakage, spend anomalies management, etc. By understanding where and how money is being spent, organizations can make informed decisions about where to cut costs. Spend analytics can be used to find both short-term and long-term savings opportunities. In the short term, spend analytics can help to identify one-time savings, such as by negotiating better prices with suppliers. In the long term, spend analytics can help to develop strategic sourcing plans that will result in sustained cost savings. As part of spend analytics, the following are some of the key aspects of spend & savings from which meaningful patterns can be discovered, interpreted, and communicated:

• What is the product or product category which is bought?
• Which spend category would the product be mapped to?
• From which suppliers, the products are brought?
• How much is spent on the bought products?
• Who (business units, business divisions across different regions) bought the products?
• What are the challenges related to each spend category? Some of the common challenges related to spend are tail-spend management, price deviations, volume bundling across different business unit / locations, etc?
• What are the opportunities for cost savings and how can the savings be maximized?

The following are the different classes of users who will be dealing with spend analytics for driving their decisions:

• Category managers, sourcing managers, supplier relationship managers from the procurement team
• Regions & countries’ leadership team
• Business/finance team
• Procurement executive leadership

The following represents some of the questions which can be answered using spend analytics:

• What are the categories & who are the suppliers where major spends (more than 70%) are happening?
• What are the categories where maximum and minimal spends is happening?
• Can the suppliers be consolidated in relation to tail spend?
• How is spend & savings split between preferred suppliers and others?
• What percentage of spend are direct and indirect spends?
• What percentage of spend is through contract and non-contract?
• What is the threshold above which procurement manage the spend? Would the threshold need to vary based on regions & countries?
• How is my spend looking across different regions and countries?
• How many POs have been raised?
• Can the spend be forecasted for one or more categories of spends?
• Can the spend be classified as maverick spending or otherwise?
• What are savings opportunities?
• What are the instances of spend anomalies?
• What percentage of spend across different categories is managed through catalog buying?

Apart from the above, data scientists can come up with different hypotheses and work with data visualization experts and analytics engineers to create appropriate dashboards and advanced analytics solutions. The primary objective for the category managers is to maximize the cost savings while creating innovative data products. The following are some of the solution approaches that can be used to achieve this objective:

• Explore the option of eAuction: eAuction is a spend management tool that helps organizations to effectively and efficiently manage their spend using online marketplace technologies. eAuction can help organizations to reduce the cost of their Spend by allowing them to drive down the prices of goods and services through competitive bidding.
• Payment terms optimization: Payment terms optimization is the process of reviewing and adjusting the timing and conditions of payments to suppliers in order to improve a company’s working capital position. The goal of payment terms optimization is to optimize a company’s spend while maintaining its supplier relationships. By analyzing data on past payments, companies can identify patterns and trends that can be used to negotiating more favorable payment terms with suppliers. Payment terms optimization can be a complex process, but the potential benefits make it worth the effort. Payment terms optimization can lead to improved working capital management, increased savings, and enhanced supplier relations.
• Volume discounts from suppliers
• Suppliers’ consolidation resulting in discounts
• Optimal demand management / inventory planning resulting in better spending and increased savings
• Prevention of maverick spends
• Tail-spend management
• Cost bundling of items purchased across different regions, BU, etc

Why Spend Analytics?

Spend analytics is very strategic to procurement analytics because of the following reasons:

1. Savings opportunities finder: Help identify savings opportunities with the help of ABC analysis by identifying spend categories where 80% of spend is made. It is with these suppliers with whom negotiations are done to carve out savings opportunities.
2. Tail spend analysis: Help analyze spends in different categories with the help of ABC analysis thereby identifying processes related to spend areas (tail spend) that can be optimized for greater operational efficiencies.
3. Supplier relationship management (SRM): Help analyze suppliers with whom spends are made. These suppliers can be further categorized into strategic and non-strategic suppliers. This helps in appropriately maintaining supplier relationships. In addition, it helps in consolidating spends with fewer suppliers in case of tail spend thereby leveraging relationships to fetch greater discounts.
4. Spend analysis by regions and countries: Provides insights on spend by regions and countries. This helps in tracking spends and detecting anomalies if any.
5. Avoid spend leakage: Avoid spend leakage in terms of avoiding spend in the categories which are mandated to be avoided.
6. Internalize spends: Identify spend which can be internalized. That would result in cost avoidance
7. Volume bundling: Identify opportunities of bundling spends by analysing spend of same items across different BU / regions at different costs.

Spend Analytics – Descriptive & Predictive Use Cases

The following are different use cases of spend analytics including descriptive and predictive analytics which can drive decision making in relation to spending:

• Descriptive spend analytics: Descriptive spend analytics can be used to extract hindsight (what has happened in the past?) in relation to the following areas:
  • Perform analysis in relation to spends in different categories to understand the major spend areas using different KPIs and suppliers with whom these spends happened. This can be used to identify strategic or preferred suppliers where effort can be put in to achieve cost savings.
  • As part of performing spend analysis, identify categories, where tail spend, happens (5% of spends). This can be used to achieve operational efficiencies as 70-75% of the effort is traditionally found to be put with tail spends. If these efforts can be put into strategic spends, one can achieve greater cost savings. One can apply machine learning-based clustering techniques to identify spends belonging to tail spends along with tail spends which can be steered to a few preferred suppliers rather than multiple suppliers.
• Predictive spend analytics: The following can be different predictive analytics use cases related to spend analytics:
  • One can use it to predict spends (spend forecasting) in different categories which could be further used for budget planning at regular interval and most importantly in the start of the year. Machine learning algorithms in relation to time-series forecasting can be used to predict the spends.
  • In addition to spend forecasting, another important use case where machine learning can be used is to classify the spend as maverick spend or otherwise.
  • Another use case is to classify whether the spend is associated with the right commodity codes (UNSPSC). This can be very helpful in ensuring that spends are booked with the right commodity codes for audit purposes.

Some Popular Spend Analytics Products

There are a number of software platforms that offer spend analytics capabilities. These platforms typically collect data from a variety of sources, including financial transaction data, invoices, purchase orders, and receipts. They then use this data to generate reports that allow organizations to see where they are spending their money and identify potential areas for cost savings. The following is a list of some of the popular spend analytics products:

• Orpheus
• GEP
• Sievo
• Zycus