← Back to blog

Vendor Due Diligence: What It Covers and How to Run It

November 16, 2026
Vendor Due Diligence: What It Covers and How to Run It

Vendor due diligence is the structured review a company does before it trusts a third party with its data, systems or operations. For B2B software companies it runs in two directions: you vet the vendors you buy from, and your own customers vet you, usually through security questionnaires and due diligence questionnaires (DDQs). Done well, both sides share one evidence base, kept current and answered from a single source. That is how we approach it at Skypher, and it is the idea this guide follows.

The phrase also has a second meaning in M&A, where a seller commissions a report on its own business before a sale. We cover that briefly below so you can tell the two apart.

Skypher
Answer Customer Reviews More Efficiently
For software teams answering customer diligence, Skypher uses AI models to automate security questionnaires and supports integrations across more than 30 online portals.
Visit

TL;DR

  • Vendor due diligence is a pre-contract assessment of a third party's financial, security, operational and compliance risk, repeated on a schedule after signing.
  • Scope by exposure first: tier vendors by the data they touch and how critical they are, then match the depth of review to the tier.
  • Ask for current evidence (SOC 2 report, ISO certificate, penetration test summary, data flow diagram), not just questionnaire answers.
  • If you sell software, every customer's diligence lands on your security team, so a maintained answer library and a trust center shorten the cycle.
  • In M&A, "vendor due diligence" means something different: a seller-commissioned report for buyers.

Table of Contents

What vendor due diligence means for a security or GRC team

In third-party risk management (TPRM), vendor due diligence is the work of gathering and judging evidence about a supplier before you sign, so you know what risk you are accepting. It covers more than cybersecurity: financial stability, legal and reputational history, operational resilience, privacy, and the vendor's own supply chain.

Regulators treat it as a life-cycle activity rather than a one-time gate. The 2023 interagency guidance on third-party relationships, issued by the Federal Reserve Board, FDIC and OCC, describes risk management for all stages of a third-party relationship and says it should reflect the level of risk, complexity and size of the banking organization and the nature of the relationship. That proportionality principle is the practical heart of the topic: the depth of your review should follow the risk the vendor introduces.

The same logic appears outside banking. NIST's guidance on cybersecurity supply chain risk management tells enterprises to perform due diligence on their specific requirements to judge whether an IT product or service is fit for purpose, because commercial products are built for a general market rather than your threat environment.

If your team is still building the wider program around this, our overview of third-party risk management for tech and finance professionals lays out the foundations.

The two directions of vendor due diligence

Most guides describe only the buying side. If you work at a B2B software company, you live on both.

DirectionWho runs itTypical artifactsMain pressure
Buying side (you assess a vendor)Security, GRC, procurement, legalIntake form, tiering, DDQ, SOC 2 report, contract clausesApproving the right vendors without stalling the business
Selling side (a customer assesses you)Security, solutions engineering, salesSecurity questionnaire, DDQ, trust documentation, call with your security leadClosing deals while hundreds of questions queue up

The second row is where deals slow down. A prospect's diligence request often arrives mid-cycle, in a spreadsheet, a portal or a PDF, and it is usually routed to the same few people who run your actual security program. Later sections show how to shorten that loop.

Step 1: Tier your vendors before you assess them

The most common mistake is giving every vendor the same questionnaire. A design-tool plugin and your cloud data warehouse do not deserve identical scrutiny. Separate two ideas:

  • Inherent risk is the exposure the relationship creates before any controls are considered: the type of data shared, how critical the service is to your operations, whether the vendor processes transactions or has privileged access.
  • Residual risk is what remains after you review the vendor's controls and certifications.

Start with inherent risk, because it decides how deep the review goes. Practical tiering questions include:

  1. Will the vendor store, process or transmit regulated or sensitive data (personal data, health data, payment data, credentials)?
  2. Will it have privileged or network access to your environment?
  3. Does a core workflow depend on it, so that an outage hurts customers?
  4. Are there data residency or cross-border transfer requirements?

A typical result is three tiers: critical (full review, contract security terms, annual reassessment), standard (a shorter questionnaire and a report review) and low (a light intake check). Document the tiering logic so an auditor can see why a vendor landed where it did.

Step 2: Collect the information that matters

For software and AI vendors, an enterprise vendor due diligence checklist from Apptension lists the items worth covering. Adapted for a security team, the evidence request looks like this:

  • Data types and data flow: what enters the product (PII, PHI, payment data, credentials, model prompts), where it is stored, and how it leaves.
  • Hosting: cloud provider and region, and whether the environment is single-tenant or multi-tenant.
  • Integration: SSO, SCIM, API keys and the network paths into your environment.
  • Subprocessors: the list, what each does, and which data it touches.
  • Go-live plan: timeline and a rollback plan if the rollout goes wrong.

On top of that, ask for independent assurance. The strongest single document for most SaaS vendors is a SOC 2 report. Insist on understanding whether it is Type 1 (controls at a point in time) or Type 2 (controls tested over a period), and read the exceptions and the scope, not just the cover letter. Our explainers on reading SOC attestations and asking for Type 2 and on what a SOC 2 sample report shows walk through what to look for.

Also request an ISO 27001 certificate where relevant, a penetration test summary, a business continuity and incident response overview, and breach history. Flag evidence older than 12 months for refresh.

Go beyond security

Security is one slice. A complete file also covers financial health, litigation, sanctions and reputational exposure, concentration risk and operational resilience. If a vendor is secure but fragile financially, you still carry the risk. Our piece on vendor risk assessment beyond security shows how to widen the lens without widening the workload for every vendor.

Step 3: Review, score and decide

Once the vendor responds, a reviewer compares answers with the evidence. Answers that say "yes" without a supporting document deserve a follow-up. Score findings against the tier: a gap that is tolerable for a low-tier vendor can be a blocker for a critical one.

Three outcomes are typical: approve, reject, or approve with conditions. Conditions should be specific and dated, for example closing a named finding within a set period, adding a breach notification window to the contract, or restricting the data the vendor may receive.

Breach notification terms deserve their own attention. If your own regulatory obligations require you to notify within a short window, a vendor that commits to a longer one creates a gap you must close in the contract, not in a footnote.

Step 4: Put the findings into the contract

Diligence that never reaches the contract does not protect you. Translate findings into terms: security requirements, audit and evidence rights, subprocessor change notice, incident notification timing, data return and deletion, and termination rights. The guidance cited above is explicit that accountability does not transfer with the work: in banking, the institution remains responsible for managing the risk of its third parties.

For AI-enabled vendors, add model-specific controls: how prompts and outputs are retained, whether your data trains shared models, and what the vendor commits to under emerging rules. This breakdown of contract controls for AI vendor risk is a useful companion when the vendor you are reviewing ships AI features.

Step 5: Monitor, and know when to repeat diligence

Approval is the start of the relationship, not the end. A sensible cadence:

  • Critical vendors: reassess annually and review their new SOC 2 or ISO evidence as it is issued.
  • Standard vendors: reassess every one to two years, or at renewal.
  • Event-driven reviews: repeat the relevant parts of diligence when something changes. Typical triggers are a security incident at the vendor, a change in ownership, a new data type or use case on your side, a new subprocessor, or a significant expansion of access.

The interagency guidance frames oversight as ongoing across the relationship's life cycle, so a calendar-only approach misses events between reviews. Build the triggers into your intake and procurement workflow so a business owner expanding a vendor's scope has to tell the risk team.

Our guide to implementing vendor management programs and the third-party risk assessment tips for compliance professionals cover how to structure the recurring work.

When you are the vendor: answering diligence faster

If you sell software to mid-size and large organizations, you receive the other side of this process every week. The problems are consistent: questionnaires arrive in different formats, the same question is worded twenty ways, answers live in policies, past questionnaires and the heads of two or three people, and a deal waits for them.

A workable approach has four parts:

  1. Build one source of truth. Keep policies, certifications, past answers and architecture documents in one place with version history, so an answer can be traced to a current document.
  2. Answer from that source, not from memory. Retrieval-based AI pulls the relevant passage from your own documents and drafts the response; a confidence score tells reviewers which answers need a human look first.
  3. Collaborate where people already work. Subject matter experts answer in Slack or Microsoft Teams rather than in a spreadsheet they have to find.
  4. Publish what customers ask for most. A trust center that exposes your certifications, policies and reports on request lets many prospects self-serve before a questionnaire is even sent.

This is the problem Skypher was built around, and the next section shows how it maps to each of these parts. If you want the buyer-side angle on what customers expect from vendors, see our 2026 guide to vendor cybersecurity for tech and finance teams.

How Skypher fits into vendor due diligence

Our Questionnaire Automation Tool parses every questionnaire format with proprietary AI models and returns answers with an AI confidence score, so reviewers can spend their time on the answers that need judgment. It can answer 200 questions in under a minute, and it connects to online portals through 30+ API connectors, including OneTrust and ServiceNow, so responses reach the customer's system without copy-and-paste.

Content comes from where your knowledge already sits: Confluence, Notion, Google Drive, OneDrive and SharePoint, with content management and document versioning to keep answers current. Collaboration is real-time and open to unlimited users, with Slack and Microsoft Teams integrations, including chatbot support. A customizable Trust Center lets customers view your security posture before they ask, and our DDQ automation handles the longer, fiduciary-style due diligence questionnaires that investors and financial institutions send.

The Questionnaire Automation Tool is the place to start for most teams, and the Trust Center suits companies that want to reduce incoming requests. If due diligence questionnaires from investors and financial institutions are your bottleneck, see DDQ automation. Pricing is quote-based and depends on usage and features; the pricing page explains how to get a quote.

Skypher

Vendor due diligence in M&A: a different meaning

If you arrived here because of a deal, note that in mergers and acquisitions the term means something else. There, vendor due diligence (VDD, or sell-side due diligence) is an independent review of the business being sold, commissioned and paid for by the seller before going to market. According to CTA Acquisitions, the seller pays for it, it is done before marketing and it is shared with serious bidders, whereas buy-side diligence is paid for by one buyer, done after a letter of intent, and kept private.

The seller pays, but the report is only useful if buyers trust it. PwC Legal Belgium explains that in a legal VDD the adviser may owe a contractual duty to buyers who rely on the report, and is expected to assess material issues impartially. Even so, buyers still run confirmatory diligence; a VDD report gives them a head start and narrows what they need to dig into.

Tech-specific questions still apply in a deal: a buyer will examine the target's security posture, certifications and customer questionnaires. Companies that already keep their evidence organized for customer diligence tend to find that part of a sale easier to document.

A worked example: tiering and answering in one week

Imagine a 300-person software company with 120 vendors. The GRC lead tiers them in a spreadsheet using the four questions above and finds 9 critical vendors, 31 standard and 80 low. The 9 critical vendors get a full evidence request, with SOC 2 Type 2 reports read for scope and exceptions, and contract clauses for notification and subprocessors. The 31 standard vendors get a shorter questionnaire. The 80 low-tier vendors get an intake form and a sanity check on data shared.

Vendor tiers and their review requirements

At the same time, the company's own sales team has three enterprise prospects each sending a 250-question security questionnaire. Without a shared answer library, each one pulls an engineer off project work for days. With one, the first pass is generated from existing documents, the security lead reviews low-confidence answers, and the sales director sees a status rather than chasing. The time saved goes back into the vendor reviews above, which is how the two directions of diligence support each other.

Common mistakes to avoid

  • One questionnaire for every vendor. It wastes effort on low-risk vendors and thins out attention on critical ones.
  • Accepting answers without evidence. A "yes" to an encryption question is not the same as a report that tests it.
  • Treating approval as final. Without triggers and a schedule, the file ages while the vendor changes.
  • Ignoring the contract. Findings that never become terms cannot be enforced.
  • Letting answers go stale on the selling side. An answer library that is not versioned will eventually send a customer last year's certification status.

Our perspective

We build for the people on the receiving end of due diligence, and what we hear again and again is that the bottleneck is rarely knowledge. Security teams usually know the answer. The cost is the hours spent finding it, rewording it and getting it reviewed, for the tenth time this quarter.

That is why we designed Skypher around your own documents and a retrieval approach that drafts each answer from them, with a confidence score so reviewers know where to look. We think the best diligence experience for a customer is one where the evidence is already visible, current and consistent, whether that is through a completed questionnaire or a Trust Center they can browse.

Buying-side and selling-side diligence are the same discipline viewed from two chairs. A team that keeps its own evidence in order is a better judge of what to ask others for, and a better partner when a customer asks the same of them.

Gaspard

FAQ

What is vendor due diligence?

It is the process of gathering and assessing evidence about a vendor, supplier or other third party to understand the risk of working with them. It covers security, financial, legal, operational and compliance factors, and it starts before the contract and continues afterward. In M&A the same words describe a seller-commissioned report on the business being sold.

What is the difference between buyer and vendor due diligence?

In an acquisition, buyer diligence is commissioned by one buyer after a letter of intent and stays private, while vendor due diligence is commissioned by the seller before marketing and shared with serious bidders. In third-party risk management, the buying company runs due diligence on its vendors, and the vendor answers it.

What should a vendor due diligence checklist include?

Cover company information and financial health, security evidence (SOC 2 report, ISO certificate, penetration test summary), data types and flows, hosting, integrations, subprocessors, incident history, business continuity and litigation. Scale the depth to the vendor's tier and flag evidence older than 12 months.

How often should vendor due diligence be repeated?

Reassess critical vendors at least annually and standard vendors at renewal or every one to two years. Repeat the relevant checks sooner after a security incident, an ownership change, a new data type or a new subprocessor.

How can a software company respond to customer due diligence faster?

Keep your policies, certifications and past answers in one versioned source, draft responses from it with AI, and have reviewers check the lowest-confidence answers first. Publishing a trust center also lets customers find common documents themselves.

Sources