Why Most Third-Party Risk Programs Are Taking The Wrong Measurements

Direct Source Verification: This story is aggregated from Forbes (forbes.com). Full reporting rights and copyright belong to the primary publisher.
Girish Redekar is Co-founder and CEO of Sprinto, building the future of trust for high-growth companies.

Girish Redekar is Co-founder and CEO of Sprinto, building the future of trust for high-growth companies.

getty​I have answered security questionnaires as a vendor and sent them out as a buyer. On both sides, they do something useful: They surface what you cannot learn about a company from the outside.

The trouble starts when the questionnaire becomes the risk program instead of being one input to it.

A vendor tells you they enforce MFA, patch critical vulnerabilities within a defined SLA and encrypt customer data. All useful. But none of those answers tells you whether privileged access has drifted since. Or that the vendor shipped an AI feature last Tuesday, and your customer data now passes through a model provider you have never reviewed.

Every one of those answers can still be accurate. The questionnaire tells you what the vendor’s controls were designed to do. Your risk program needs to know what is actually happening today.

Measuring vendor risk only through questionnaires has a well-known flaw: The vendor grades themselves. But there are two other flaws that do more damage.

The first is age. A questionnaire is like the fire safety certificate framed in a restaurant lobby: accurate on the day it was issued. If you review vendors once a year, that snapshot guides your decisions for the next 12 months.

The second is that very different vendors often get pushed through versions of the same process. The payroll provider with salary data for your entire company and the tool your design team uses for designing mockups can end up answering much of the same questionnaire and landing in the same folder marked approved.

Some teams respond by sending the questionnaire every quarter instead of every year. I understand the instinct, but it gets them the same snapshot four times as often. AI makes this even easier, with vendors using it to fill out questionnaires and buyers using it to grade them. The paperwork moves faster, but what changed after approval is still invisible.

Imagine a vendor passes review in March. In August, your sales team connects it to the CRM, and it starts holding customer records. Nobody reviews the change because the vendor is already approved. In November, the vendor discloses a breach, and you find out your customer data was part of it.

The analyst who approved the vendor can show you a complete file. Her work ended at approval. The eight months after it fell between teams.

A questionnaire tells you what a vendor says should be true. You also need evidence of what is changing around that vendor. That is what I mean by exposure evidence.

Most of that evidence comes from how the business runs after approval. A tool approved for marketing gets rolled out to sales and starts holding pipeline data. The vendor switches on an AI feature by default in a routine update. Neither change passes back through procurement, and an external security score could miss both because nothing visible from the outside changed.

Continuous monitoring is what makes those signals useful. It watches for changes that alter the risk you originally accepted: a broader permission, a new use case, more sensitive data or access that should have expired. It also tells you which changes need attention and which do not, so you are not sending every vendor back through the same review.

All of it serves one decision: If this vendor came up for approval today, with what you now know, would you still approve it?

A board member asks a simple question: If one of our vendors were breached tomorrow, how exposed are we?

You might refer to the approval status: Every critical vendor was reviewed this year, and every one passed. That is a true statement about last spring.

A better answer sounds like this: Two vendors expanded into systems holding customer data this quarter, and both were re-reviewed. One switched on an AI feature that sends data to an outside model provider, so we reopened the assessment and restricted its use while we review the change. Here is what we are watching next quarter.

The second answer takes three minutes and gives the board something to act on.

If I had to replace questionnaire completion rates with a single number, it would be the days between a change at a critical vendor and someone on your team knowing about it. On an annual review cycle, that can be months.

You can start shrinking that number this week. List the 10 vendors that hold your customer data or touch production, and name one person who owns changes to each. Then ask that person what has changed at each vendor since it was last approved. If they have to email the vendor to find out, that gap is your answer.

The certificate in the lobby will keep saying the restaurant passed inspection. It won’t tell you someone stacked boxes in front of the fire exit last week. Your program should be measuring how quickly you notice.​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

Original Source
https://www.forbes.com/councils/forbestechcouncil/2026/11/20/why-most-third-party-risk-programs-are-taking-the-wrong-measurements/
Visit Forbes ↗
SHARE STORY:
𝕏 f in

Related Coverage in Business