Think Global, Buy Local: Saving money using AI to support local purchasing
CONTENTS
- Summary
- Background
- Local Expenses
- Practical AI
- Case Study
- ‘Making it Happen’
- Conclusion
ABOUT THE AUTHORS
Scott Setrakian - Vice-Chairman, Foundry.ai
Scott leads Foundry.ai’s San Francisco office. Prior to joining Foundry.ai, Scott was co-founder and Managing Director of Applied Predictive Technologies. Previously, Scott sat on the Board of Directors of Mercer Management Consulting and ran the firm’s global Oil, Gas, Chemicals, Pharmaceuticals and Process Industries Consulting Group. Scott received an MBA and an A.B. in Human Biology from Stanford University. He sits on the Board of Directors of the Buena Vista Funds, the William Saroyan Foundation and the San Francisco Zoo.
Venu M. Amar - COO, Supplier.ai (by Foundry.ai)
Venu is a Vice President at Supplier.ai. Prior to joining Supplier.ai, Venu was at Nike, where he designed and led the integration of Zodiac, an AI startup the firm acquired in 2018. Previously, Venu was a Vice President at Zodiac and a Principal at Applied Predictive Technologies. Venu received a B.S. in Economics, summa cum laude, from the Wharton School at the University of Pennsylvania.
SUMMARY
Every company knows that effectively purchasing the goods and services necessary to run its day-to-day business has a significant impact on profitability. To this end, most corporations manage their largest expenses with dedicated professionals and standardized processes.
However, companies with extended physical networks such as retailers, hotels, REITs, restaurants, and franchise operators have a special challenge: managing the myriad expenses that occur locally across their geographic footprint. These expenses cannot be cost effectively managed by a central professional group and, as a result, are often left unoptimized.
Advances in Artificial Intelligence (AI) now enable these types of companies to more effectively manage local purchases. These ‘Practical AI’ solutions drive down total local expenses by about 10%, on average, without requiring significant management time or sacrificing quality. In doing so, they can drive millions of dollars to the bottom line.
BACKGROUND
For good reason, companies devote significant attention to their large-scale purchases to ensure that they are getting the best prices, terms, and service from their suppliers. These purchases are usually transacted by a dedicated buying group that — consistent with long-held best practices — will contact multiple potential suppliers, qualify them through a process requesting information, and manage a bidding process meant to elicit each contender’s best quote.
This process is effective, but it takes significant work to properly execute. As a result, there is a volume threshold that represents the lower limit of the cost/benefit relationship. Simply put, investing $25,000 of resource time to save 10% on a $1 million purchase is well worth it, but investing $25,000 to save 10% on a $100,000 purchase is not.
Companies with extended physical networks such as retailers, hotels, REITs, restaurants, and franchise operators have a large proportion of smaller purchases due to the geographic distribution of their locations. They also have ongoing needs for locally-provided goods and services that make up a significant number of transactions and a meaningful amount of money: up to 30% of a company’s buying transactions, collectively representing a material portion of total operating expenses.
Managing these expenses in support of individual units or districts can be complicated and inefficient. And using dedicated professionals to do so is unattractive — it takes too much time and costs too much, compared to the benefits of improving any single transaction. However, ongoing advancements in AI have opened up opportunities to apply purchasing best practices to these smaller and more frequent expenses — reducing costs while maintaining or improving quality.
“With numerous locations spread across many geographies, multi-unit operators face the unique challenge of managing the purchases of a wide variety of goods and services that are delivered locally.”
LOCAL EXPENSES
With numerous locations spread across many geographies, multi-unit operators face the unique challenge of managing the purchases of a wide variety of goods and services that are delivered locally. Some of the most common purchase categories are outlined below (Figure 1).
On balance, local expenses constitute tens-to-hundreds of millions of dollars in annual spend for these types of companies. Accordingly, paying the right prices can send many millions per year straight to the bottom line.
However, given the broad geographic distribution and sheer volume of these expenses, too often, senior operators have lamented that in lieu of a rigorous process, the local manager “just selects the supplier that lives in his or her neighborhood.”
Nonetheless, the stakes are far from theoretical. Figure 2 highlights Supplier.ai’s analysis of several thousands of purchases totaling hundreds of millions of dollars in spend. We see that as the number of bidders increases for a given transaction, the ultimate price decreases.
The biggest improvements are recognized when the number of bidders moves from just one or two, to three or more — and for the majority of these transactions, only one potential supplier was contacted.
Additionally, adding multiple rounds to the bidding process can further reduce prices. Figure 3 shows the progression of quotes that three suppliers provided when asked to lower their prices in order to successfully compete for the business.
And, field work in support of local purchases indicates that this relationship is, if anything, magnified in these settings given the wide range of quotes providers often supply (Figure 4).
Reaping the benefits of discovering and securing the lowest quality-adjusted price for local expenditures is a textbook application for ‘Practical AI.’
PRACTICAL AI
AI has recently surpassed ‘Big Data’ as today’s leading business buzz-phrase. One reason AI is becoming such a focal point is that computing and data management capabilities have recently crossed a threshold where, in many cases, software can do what we humans would do if they had unlimited time, access to complete data, and computational infallibility — all at a very low cost.
In popular culture, AI is robots playing Jeopardy! and paintings created by algorithms. These are fun to think about, but the greatest tangible value being created by AI today is much more prosaic: software powered by the application of data and math that is designed to statistically improve key business processes. We call these software tools ‘Practical AI.’ Practical AI is developed to solve business problems in ways that are simultaneously better, faster, and less expensive than human alternatives.
There are three specific characteristics within business processes that qualify them for potential improvement by Practical AI:
- Repeatable decisions that occur at high volume
- Decisions that, in aggregate, have high profit leverage
- Decisions with large data sets available that are not fully optimized
Each of these characteristics applies directly to local purchases, and AI software developers have accordingly attacked the opportunity. They are deploying a combination of natural language processing, machine learning, machine vision, and large-scale multi-source data integration to power tools that address the variety of challenges associated with these types of purchases.
Specifically, the AI-enabled system takes input of the specific purchase request, automatically identifies multiple prospective suppliers, and manages the process of securing the best vendor based on price, quality, and relevance. The process follows best-practice protocols, while absorbing only a minimal amount of management time. When repeated thousands of times a year, the results are significantly lower costs and tangible improvements in profitability.
CASE STUDY
- Scenario -
- Solution -
‘MAKING IT HAPPEN’
As with any procedural change, it is important that the company communicate, support, and normalize the new process for local purchases. We’ve observed that three discrete management tactics will significantly increase the likelihood for success:
- Participants will need up-front communication, initial system orientation, and ongoing availability of user support. The AI systems that support local purchases are intuitive and easy to use but require appropriate introduction. Little will transpire without a good start. These communications can be organized by phone and/or over the Internet.
- Senior support is vital. One CFO has put into place a rule that no purchase above $2,500 will be approved without at least three bids, and many other companies follow analogous protocols.
- Tracking results is straightforward and can build enthusiasm & support. Additionally, sharing pricing across the network by function can help inform in-market purchase decisions.
CONCLUSION
Companies with extended physical networks such as retailers, hotels, REITs, restaurants, and franchise operators have the opportunity to optimize tens-to-hundreds of millions of dollars of local spend. Advances in AI have fundamentally changed the game, and the organizations that are at the forefront are ‘playing ball.’ In so doing, they are seeing results almost immediately and are saving many millions per year.
— —
