Skip to content
Resources · Blog

Technical Due Diligence Checklist for Investors 

A technical due diligence checklist covers code, architecture, security, cloud cost, key-person risk and IP licenses. A focused review takes two to four weeks.

November 21, 2026 7 min read
SecurityArchitectureBuyer's Guide

A technical due diligence checklist covers seven areas: code quality, architecture, security, infrastructure cost, the team and key-person risk, ownership of the code and its open-source licenses, and the delivery process that keeps all of it working. For a company with one main product, a focused review takes two to four weeks and should end in a written list of risks, each with an estimated cost to fix and a view on whether it changes the price or the deal terms. Investors and acquirers commission these reviews to answer one practical question: can this software carry the business plan they are paying for, without a rebuild nobody has priced? Every codebase that has shipped to customers has shortcuts in it, and the useful output separates the ones that are normal from the ones that will cost money in the first year after closing. We run reviews of this kind at RNDSOL as part of our cybersecurity and compliance work, and the checklist below is the one we work from.

What Does a Technical Due Diligence Review Cover?

  • Code quality: read access to every repository, test coverage around the core logic, the age of the main dependencies, and which parts changed most in the last six months
  • Architecture: a diagram that matches what is deployed, how the system scales with customers, single points of failure, and the parts the team avoids touching
  • Security: the last penetration test and what was fixed, secrets handling, access control, encryption, logging, and any incidents in the past three years
  • Infrastructure cost: twelve months of cloud invoices, cost per customer or per transaction, and how that cost moves as the business grows
  • Team and key-person risk: who wrote and who understands each major component, the split between employees and contractors, notice periods, and documentation
  • IP and licenses: assignment agreements from every employee and contractor who wrote code, and a scan of open-source dependencies and their licenses
  • Delivery process: how code reaches production, how often, how failures are detected, and how long a restore from backup takes
  • Compliance: the data protection and sector rules that apply to the company's customers and markets, and the evidence that it meets them

How Do You Assess Code Quality and Architecture?

Start with the version control history, which shows how many people commit, which modules change constantly, and whether one name is on most of the commits. Then build and run the system from the repository using only the written instructions; if a senior engineer cannot get it running in a day, onboarding a new team after closing will be slow and expensive. Read the parts that hold the business logic, such as billing, pricing or the core workflow, rather than sampling files at random. Look for automated tests around that logic, dependencies that are several major versions behind, and any framework or runtime that has reached end of life, because each of those becomes a priced upgrade. On architecture, compare the diagram the team presents with what runs in the cloud account. Ask what breaks at ten times the current load, and where the answer is a single database or a single server, estimate what separating it would take. A system that serves today's customers well but needs restructuring to serve the growth plan is common, and it is a cost to put in the model, not a reason to walk away. Our architecture decisions playbook covers the trade-offs a reviewer should expect the team to explain.

Security and the Cost of Running the Product

Security findings matter most when they touch customer data or money. Ask for the last penetration test report and the list of fixes made after it, then check a few of those fixes yourself. Look at how secrets are stored (in a vault or the cloud provider's secret manager, or in the repository), whether production access is limited to named people with multi-factor authentication, and whether backups are encrypted and restore-tested. Then read the cloud bills. Divide twelve months of infrastructure spend by customers or transactions and watch the trend: a cost per customer that falls as the company grows is healthy, while one that rises usually means an architecture that scales by adding money. It is common to find a share of spend going to idle environments, oversized databases and logs nobody reads, which is upside for the buyer. Our cloud cost optimization checklist lists the usual places it sits.

Who Owns the Code, and Are the Open-Source Licenses Safe?

Ownership is the area most likely to become a legal condition of the deal. Every employee and contractor who wrote code should have signed an agreement assigning their work to the company, and any agency that built an early version should have a contract that transferred the IP on payment. Gaps are common in startups that began with freelancers, and are usually closed before signing with retrospective assignments. For open source, run a software composition analysis tool across every repository to produce a list of dependencies and their licenses. Permissive licenses such as MIT, BSD and Apache 2.0 rarely cause problems. Copyleft licenses need a closer look: the GPL can require source code to be released when the software is distributed, and the AGPL can extend that obligation to modified versions that users interact with over a network, which matters for a SaaS product. Whether a given use triggers those obligations depends on how the code is combined and shipped, so treat the scan as a list of questions for counsel, not a verdict. Check, too, for copied code with no license at all, and for paid components whose licenses do not transfer on a change of control.

What Is Key-Person Risk in a Software Company?

Key-person risk is the chance that the product stalls when one person leaves, and in small engineering teams it is one of the most common serious findings we see. The signs are easy to measure: one engineer behind most commits to the billing or core module, production credentials held by one person, and deployment steps that live in someone's head. Interview the technical leads separately and ask each one to explain a component they did not write. If nobody can, that component is held by one person. Check whether the key people are employees with notice periods and equity that keeps vesting after closing, or contractors who can leave with a week's notice. The fixes are known and cheap compared with the risk: retention terms for two or three named people, a written runbook, shared ownership of credentials, and a few months of a second engineer pairing on each critical module. Put those in the deal terms or the 100-day plan rather than hoping they happen.

How Long Does a Technical Due Diligence Review Take?

A focused review of one product with a team of up to about 15 engineers takes two to four weeks from the day access is granted, and access usually takes longer to arrange than buyers expect, so request it in the first week of the process. A light review, suited to an early-stage investment, is one senior engineer for one to two weeks: repository access, cloud bills, the last security report and interviews with the technical leads. At $1,400 to $2,200 per engineer-week, that is $1,400 to $4,400. A full review before an acquisition, with a security assessment, a license scan, an infrastructure cost model and a hands-on build, is typically two engineers for two to three weeks, or $5,600 to $13,200 at the same rates. The output should be a short report listing each risk, its severity, the evidence, the estimated cost to fix in engineer-weeks, and whether it should affect price, warranties or the 100-day plan. Our post on software maintenance cost gives planning figures for what the product will need each year after closing.

The Technical Due Diligence Checklist

  • Read access to every repository, the cloud accounts and the CI pipeline, granted in week one
  • A working build from the repository using only the written instructions
  • Test coverage and dependency age for the core business logic
  • An architecture diagram checked against what is deployed, with known scaling limits
  • The last penetration test, the fixes made since, and any incidents in three years
  • Secrets, production access, backups and restore tests reviewed
  • Twelve months of infrastructure cost, expressed per customer or per transaction
  • A commit history review showing who owns each critical module
  • Retention or handover terms for any engineer who is a single point of knowledge
  • Signed IP assignments from every employee, contractor and agency that wrote code
  • An open-source license scan, with copyleft and unlicensed code referred to counsel
  • A risk list with a cost to fix for every finding, tied to price, warranties or the 100-day plan

Have a project that needs this kind of thinking applied to it?