In procurement, critical knowledge often lives in inboxes, spreadsheets, or people’s heads. When those people leave, so does the context behind past decisions and supplier relationships. In this Q&A, Spencer Penn, CEO and co-founder of LightSource, explains why institutional memory is one of procurement’s most undervalued assets, and what happens when companies finally start capturing it by design.

What’s Related

Supply Chain 24/7: Why does institutional memory break down so easily inside supply chain and procurement teams?

Spencer Penn: Company information can live in one of two places: systems or people’s heads. Institutional memory is the tacit knowledge within a business. At its core, there’s a dichotomy between the two places where information lives. The first is explicit knowledge (the hard information codified in systems or process documentation), and the other is tacit or “tribal” knowledge. The difference between the two is whether the information resides in systems of record or in people’s heads.

To make this more concrete, here are some examples of tacit vs. explicit knowledge in a few key areas:

What makes tacit knowledge so ephemeral is that it’s often learned over years of personal experience, and it can be difficult to formalize, articulate, or simply transfer to others. Knowledge can range from process knowledge — “here’s how we do this” — to data — “this price should be X.”

Supply chain and procurement also have a unique distinction from other functions: it contends with networks of external parties. For example, when you think about an HR system, it’s much easier to control the data because it’s strictly internal information that the business owns. By contrast, in the supply chain, you’re dealing with the “interaction surface” between your company and a global vendor base. They all use different systems. On top of that, there are more links in the chain, including suppliers, your suppliers’ suppliers, and their raw material suppliers, often spread across multiple borders.

SC247: What usually happens when key people leave, and that knowledge leaves with them?

SP: When someone leaves the business, information leaves with them. Very rarely do people at the end of their tenure pass along everything they’ve learned. Maybe there’s a handoff document, or a few key tasks shared with a manager or direct reports, but it’s not possible, and often impractical, to transfer all the knowledge someone has accumulated in their head.

As a result, the business is left relying on limited knowledge transfer or on what has been codified in systems of record.

“When someone leaves the business, information leaves with them. Very rarely do people at the end of their tenure pass along everything they’ve learned.”

In procurement and supply chain, this can be quite embarrassing. Even at companies like Tesla, I saw teams periodically relying on suppliers for information. As prices changed from initial to final design, we would ask suppliers to provide “cost walks” — Excel files showing how design changes drove price changes. This wasn’t well codified in any internal system.

Compounding this, some organizations, such as Tesla, also experience significant turnover. When I was there in 2016, the average tenure was around 18 months. So knowledge would frequently leave the organization, with others left to pick up the pieces.

SC247: Why do so many companies still rely on inboxes, spreadsheets, or people’s memories to track critical decisions?

SP: People rely on inboxes and spreadsheets for two main reasons: (1) they’re universal and (2) they’re flexible. Everyone has an email address, and everyone knows how to use a spreadsheet. The other nice thing about spreadsheets is the flexibility – users can lay out almost any kind of data, calculation, etc., they want to store or share. However, that same flexibility is also their Achilles’ heel. They’re unstructured. There’s no data aggregation, and information becomes siloed. 

The key attributes of good systems of record are that (1) they can maintain data, (2) data entry is consistent and compliant, and (3) data can be easily retrieved. The latter is not the case with spreadsheets and emails. 

In theory, IT could go through a former employee’s inbox and dig up all the spreadsheets and data that’s relevant, but practically, this is difficult. 

SC247: You’ve said asking people to document things after the fact rarely works. Why is that such a tough habit to enforce?

SP: Documenting after the fact is hard because most enterprise systems aren’t designed to help people do the work; they’re designed to prevent people from doing the wrong thing.

Many procurement leaders believe their systems are being used to collaborate with suppliers. In practice, the real back-and-forth happens in Excel and email. Prices are negotiated, tradeoffs are made, and expectations are set outside the system.

Once the deal is done, people are asked to go back and “clean it up” in the system of record. That’s painful because the work is already finished, the context has faded, and the system wasn’t part of the workflow to begin with.

SC247: How does lost context show up in day-to-day procurement work, like negotiations or supplier relationships?

SP: Lost context shows up in very practical ways day to day:

• Not knowing which SKUs were historically negotiable versus untouchable
• Losing track of past concessions and the tradeoffs that justified them
• Forgetting how a supplier actually performed when things went wrong
• Re-evaluating alternatives that were already considered and ruled out
• Treating exceptional discounts or terms as the new baseline
• Resetting supplier expectations every time ownership changes

Critically, none of this appears in contracts or systems of record, yet it directly affects pricing, leverage, and supplier relationships.

SC247: What changes when information is captured automatically instead of relying on manual follow-up?

SP: Compliance to a system of record is an important aspect of its usefulness. If you look at a tool and see that only 50% of the available data is captured, it’s incomplete. So the real question becomes: how do you get data capture to 99%+? 

You have to either enforce rigorous processes (which are painful and tedious for people) or build systems that first make employees’ jobs faster and easier, while capturing information in the process. 

It’s the difference between forcing kids to eat vegetables, vs. giving them candy that happens to be healthy. 

SC247: Why do you see institutional memory as one of the most overlooked assets in supply chain technology?

SP: Institutional memory is one of the most overlooked assets in supply chain technology because most tools focus on recording transactions, rather than capturing the context behind decisions. Price lists, contracts, and POs are stored, but why concessions were made, which SKUs were negotiable, how suppliers performed, and what trade-offs were accepted rarely make it into the system—they stay in people’s heads.

When that context is lost, teams spend an extraordinary amount of time reconstructing history from inboxes and spreadsheets before they can even start the real work. If teams can jump straight into a software cockpit with data, history, and context already assembled, they can eliminate much of the tactical data-gathering and focus instead on strategy, risk, and supplier value creation.

The value of preparation shows up in negotiation research, too. A Harvard Business Review analysis of nearly 1,000 real negotiations across 50 countries found that the strongest negotiators succeed through deliberate preparation and data-driven strategy rather than instinct alone. On top of that, negotiators who use more open-ended questions, a hallmark of good preparation, can achieve roughly 20 % higher gains in outcomes without compromising the other party’s interests. 

SC247: Looking ahead, what does a “system of record” really need to do to support modern procurement teams?

SP: A modern system of record needs to move from being reactive to proactive.

The first wave of digitization in procurement was basic record-keeping, such as converting previously tacit information into a system. The second wave focused on compliance, using systems to prevent people from doing the wrong thing. Both were necessary, but neither actually improved teams’ performance.

The third wave is acceleration. These systems help people work faster and more effectively, and institutional memory gets captured naturally as a byproduct of doing the work, not as an after-the-fact task.

The next wave is proactive intelligence. Systems of record should continuously monitor data, signals, and changing inputs like prices, supplier performance, risk, demand,and surface what matters before it becomes a problem. Instead of asking users to hunt for information, the system should elevate the decisions, risks, and opportunities that require attention.

That’s when a system of record stops being a passive repository and starts becoming a strategic asset.

Spencer Penn is CEO and co-founder of LightSource