Procurement’s Vigilance Problem in Man and Animal
Just as animals constantly sense critical signals such as the sound of a predator approaching, procurement professionals must always be receptive to market signals that flag upstream market threats and opportunities.
This animal analogy can help us make sense of the many software solutions on offer to procurement professionals, especially in the context of how these solutions help professionals at various levels to monitor supply chains.
The Vigilance Deficit
It is very difficult to remain vigilant indefinitely. Scientific studies of people and animals confirm that the longer we spend on the watch, the less we actually see. Moreover, the longer we are watchful, the less time we can spend on other work.
In their seminal paper, The Problem of Vigilance in Animal Life, Drs. Stuart Dimond and John Lazarus describe the ways that evolution has dealt with life’s vigilance problem in animals. They offer three themes: (1) neurological adaptation; (2) learned pattern recognition; and (3) the distribution of the vigilance burden across social groupings.
In addition to helping us understand how Mother Nature addresses the need for vigilance among animals, these three themes can be used to help us classify procurement software solutions.
Neurological adaptation
Higher-order species — including humans — have developed sophisticated ways to process information. They possess relatively large brains that allocate more space for the telencephalon, the two cerebral hemispheres in advanced mammals that are responsible for sensory perception and reasoning (think ‘left brain, right brain’) (1, 2). This evolutionary investment in information processing appears to deliver some important survival advantages. For example, animals with advanced brains can sleep lightly while still remaining receptive to changes in their environment.
In the procurement profession, neurological adaptation comes in the form of software offerings that automate procedures that involve source-to-pay, procure-to-pay and supplier relationship management. These procedures, including tender management, bid management, auction execution, on-boarding, and score carding, can be tedious and attention-consuming.
Various vendors including Apex Analytics, Scanmarket, BasWare, Coupa, Ivalua, Jagger, Oracle, Siron Labs, SynerTrade, and GEP, offer these solutions. A number of niche players also compete in this space, including Achilles and Beeline for the procurement of human resources, and Birch Street Systems in the restaurants and hospitality markets.
Learned pattern recognition
Animals have the ability to quickly identify hidden opportunities. An example is pigeons searching for seeds scattered in foliage and grasses. Studies show that animals’ ability to recognize opportunities correlates with the prevalence of those opportunities in their environment. Pigeons living in an environment where sunflower seeds are common are better at identifying sunflower seeds against confusing backgrounds than are pigeons in other environments. (4, 5). It is hypothesized that this is a result of learned pattern recognition.
A class of search image software solutions provides procurement professionals with similar abilities. An example is specialty data dashboards designed to quickly explain expenditures and provide alerts. Vendors that offer these solutions include Sievo and Insight Sourcing Group. Some companies offer issue-specific search image solutions, such as RapidRatings, that dashboards the financial health of trading partners, EcoVadis that dashboards trading partner sustainability issues, and Resilinc, that evaluates and presents supplier resilience (many neurological adaptation software packages also include search image solutions.)
Distributing the vigilance burden
Remaining vigilant can be a time-consuming task, and animals in communities such as flocks share the burden with other members of the group. Scientists Murton and Lazarus measured how much time birds spend feeding as opposed to keeping watch and related these measurements to the size of each bird’s associated flock. Larger flock sizes correlated with more time feeding and less time on the watch (6).
Procurement professionals can flock together for vigilance in two basic ways. Group Purchasing Organizations including Omnia Partners, Corcentric, and CoreTrust, essentially allow companies to join large groups of buyers that consolidate a company’s spend in certain categories.
The second route procurement professionals use to join “flocks” can be called the social networking of supply bases. Some software platforms enable companies to build out a networked database of suppliers and customers. The platforms link a business to a specific sub-network of the database and customize business dashboards to the sub-network’s events (think of your favorite social media feeds). Acute threat information travels very quickly across such a network for two reasons: (1) suppliers can quickly broadcast information on disruptions to affected parties; and (2) the software vendor can monitor for threats and quickly broadcast them across their channels to relevant trading partners. Achilles, Avetta, EcoVadis, and Resilinc are among the vendors that market these “flocking” solutions.
Deus ex machina
Artificial intelligence (AI) is attracting huge interest as a technology that can transform key supply chain activities — including procurement. However, to paraphrase one of history’s great futurists, William Shakespeare, AI can be characterized as “Giving more light than heat, extinct in both.”
The expectation that software is already programmed and ready to execute strategic procurement functions is far-fetched. The real promise of AI in procurement is in the detection of upstream signals too faint and scattered for even experienced procurement professionals to reliably see. Below are a few near-term applications of AI algorithms that offer much promise.
Association Algorithms
Association rule algorithms use the energy of computers to sift through very large and tedious data sets for hidden relationships. This is the same class and application of AI as is used by banks to automatically identify potentially fraudulent charges on credit cards. As many procurement professionals know, erroneous and fraudulent invoicing can be an expensive and time-consuming problem. Similarly, procuring an item or a service can generate a multitude of disparate documents and tickets across a variety of functional teams, including purchase orders, invoices, contracts, and supplier scorecards. Intelligent matching of this documentation that digitally collates related files and looks for suspicious aberrations could be a practical and cost-effective effective AI-enabled time saver. Such solutions are discussed in publicly available documentation from Apex Archimedes and the Zycus Merlin AI Studio.
Near-Term Natural Language Processing
Natural language processing algorithms are the instructions that enable computers to infer meaning from man-made texts and speech. Because procurement is codified in written documents such as contracts and invoices, training computers to read and organize them is a potentially useful application. Vendors here offer two-step solutions: (1) document library digitization; and (2) natural language processing that combs through your company’s digitized library and automatically makes dashboards of relevant terms and conditions. These algorithms can be trained to look for discrepancies and opportunities hidden in the contract library’s fine print. Exari (now a Coupa company), Icertis, and Seal Software market products in this area. Apex Archimedes offers suites based on natural language processing that comb through company emails for discussions on payment and contract information.
Medium-Term Natural Language Processing
A logical next step, and perhaps a Holy Grail, is to automate the reading of all relevant documents and synthesize these readings into continuously updated, legally relevant procurement dashboards for agile decision making. This process would include all of the signals and dashboards mentioned above — as well as meaningful summaries of all the digital data generated by a growing array of internet-of-things sensors. These solutions would also include analyses of human-generated text broadcast by social media networks like Twitter, Instagram, and Facebook, and even insights from satellite surveillance of customers, competitors, and inventory stockpiles.
The potential for this level of automation is huge, and much work is underway to develop solutions. However, a commercially viable market for these types of solutions is some way off.
Lessons from nature
Biomimicry and industrial ecology can teach us much about how living things monitor the environment in which they live — especially in relation to how animals survive and thrive. Applying these lessons in the world of procurement helps professionals to make sense of complex man-made supply chains and to be productive while being vigilant.
