Software technical due diligence examines the technical evidence behind a product and the obligations needed to keep developing and operating it. Its scope is driven by a decision: investing in a software business, acquiring a product, taking responsibility for an application or approving a substantial next phase.
The review should distinguish demonstrated capability, unresolved assumptions and work that would be needed after the decision. It is different from buying due diligence software to manage questionnaires or customer checks. This guide focuses on assessing the software product itself. It covers technical inputs; commercial, financial and legal conclusions need their own appropriate review.
Start with the decision and its technical assumptions
Ask what must be true for the proposed next step to work. An investor might need evidence that a product can be maintained by the team expected to own it. A buyer might need to understand which services, licences and accounts the product depends on. An operating team might need to demonstrate that it can deploy and recover the release.
Turn those assumptions into questions with evidence requirements. “The platform scales” is incomplete. Identify the workload, critical paths, constraints and evidence available. If no relevant test has been run, report that gap rather than substituting an architectural diagram for measured behaviour.
Record the product version, review date, decision owner and boundary of the engagement. Clarify whether the review covers one application, several services, mobile releases, data pipelines or infrastructure. A change to the product or its dependencies may change the relevance of an earlier assessment.
Software due diligence checklist: what evidence to request
| Area | Evidence to request | Decision question |
|---|---|---|
| Product and architecture | Critical journeys, system map, service boundaries and current release inventory | What must continue working, and where does responsibility sit? |
| Code and delivery | Repositories, build process, tests, release records and reviewer access | Can the intended team verify and change the product? |
| Data and integrations | Data flows, migration responsibilities, API dependencies and external owners | Which dependencies could constrain transition or planned changes? |
| Security and operational readiness | Selected control evidence, monitoring, incident history and recovery exercises | Which important claims have been demonstrated, and which remain open? |
| Components and rights inventory | New code, existing supplier components, open-source and paid-service inventories | Which rights, restrictions and transfer arrangements need specialist review? |
| People and ownership | Maintainers, account owners, runbooks and transition responsibilities | Who can make a release or handle an incident after the decision? |
| Roadmap and remaining work | Known limitations, technical dependencies and estimates with assumptions | What technical work would the planned next phase actually require? |
Request evidence proportionate to the decision. Keep a register showing the owner, location, reviewed version and verification status of each item. A document that exists but cannot be reconciled with the current deployment should be marked accordingly.
The audit preparation worksheet can organise the code and delivery inputs. Add transaction-specific questions to the scope where relevant; the worksheet is not a complete transaction checklist or a legal opinion.
Where a code audit fits
A software code audit can examine implementation quality, selected security boundaries, tests, dependencies and delivery evidence. Technical due diligence uses those findings alongside operational, people and dependency evidence to address a broader decision.
Keep the outputs traceable. A technical reviewer might identify that a critical service has no demonstrated restore procedure. The decision owner then needs to understand who will verify restoration, what access is missing and how that affects the intended transition. The reviewer should not turn that finding into an unsupported valuation or a prediction that a deal will succeed.
NIST's Secure Software Development Framework, version 1.1 provides a vocabulary for discussing secure development practices with suppliers. It can help frame evidence questions; citing it does not establish compliance or replace assessment of the actual product.
A practical sequence for the review
- Agree questions and access. Name the decision, reviewers, baseline, exclusions and evidence that can be shared.
- Inventory the product. Reconcile repositories, deployed services, data stores, distribution accounts and important suppliers.
- Test important claims. Use agreed demonstrations, selected code review, build and test results, and operational evidence.
- Record findings and unknowns. Separate observations, interpretations, unverified claims and missing evidence.
- Review next actions. Assign owners, investigation steps and conditions for proceeding; retain the unresolved questions in the decision record.
Arrange interviews with the people who maintain the system, then check important explanations against artefacts or demonstrations. Knowledge held by one person is a dependency to understand. It is not automatically proof that the software must be replaced.
Illustrative example: acquiring a booking product
This is a hypothetical review, not an EmphaSoft transaction. A company is considering taking over a booking application. Its planned next phase assumes that a receiving team can release changes independently and continue sending confirmations through an existing email service.
The team demonstrates a build of the identified commit in a test environment. However, the email-service account is controlled by an outgoing supplier, and the review has not established whether the intended buyer can use or replace that account under the applicable arrangements.
The report should record the successful build, the unresolved account dependency, the relevant owner and the next evidence required. It can outline the technical work involved in an alternative integration, with assumptions. It should not label the dependency resolved because someone says that access will probably be transferred.
A second question concerns historical booking data. The reviewer can ask for the data model, migration process and an authorised migration rehearsal using suitable test data. The technical evidence does not by itself decide the legal basis for transferring records. Keep those responsibilities explicit.
How to read the due diligence report
A useful executive summary answers the agreed questions with evidence and confidence levels. Ask for a findings register with the reviewed baseline, observation, supporting record, affected capability, uncertainty, next action and owner. Separate repair estimates from completed validation.
Check the boundary of reassuring statements. “No issue found” should specify the paths, tests and evidence examined. “Ready for takeover” should identify the receiving team and demonstrations completed. A list of desirable engineering practices is not a record that they are implemented.
Group actions by the decision they affect: evidence needed before proceeding, work needed for the transition and improvements that can be planned later. Define the reasons for those groups. Avoid universal scores or fixed thresholds that ignore the product's actual users and operating needs.
Common gaps that deserve a follow-up question
- A diagram describes services that do not match the deployed inventory.
- The build works only on one maintainer's machine, with undocumented inputs.
- The roadmap estimate omits migration, external approvals or account dependencies.
- Backups exist, but the review has no evidence of a restoration.
- Tests pass, but do not exercise the journey underlying a key claim.
- Component or licence inventories are incomplete and require specialist follow-up.
These are review prompts, not findings about any particular product. Each needs investigation before a conclusion. Missing evidence and a verified defect should remain different statuses.
Connect technical evidence to transition planning
Use the project handover checklist to record receiving owners, deployment and recovery evidence. Use the maintenance and support guide to agree who operates the service after the transition. If substantial development follows, reconcile its assumptions with the scope and acceptance worksheet.
There is a published example of an existing-team engagement in the Sourcefit case, which lists a project audit before subsequent delivery. It does not disclose acquisition due diligence or an investment decision. Do not treat it as evidence of those activities or infer private audit conclusions from the public case.
Frequently asked questions
Is software due diligence only a security review?
No. Depending on the decision, it may include architecture, delivery capability, operational evidence, dependencies, team knowledge and technical roadmap assumptions. Security is one area whose scope should be stated.
Do we need access to production customer data?
Define the evidence needed for each question. Prefer documentation, authorised demonstrations and suitable test data where they answer it. Any access to sensitive systems or records needs explicit scope and appropriate arrangements.
Does the review tell us whether to invest?
It supplies technical evidence and uncertainties for the people making that decision. It does not replace commercial, financial or legal assessment, or guarantee an outcome.
Prepare the questions for your technical review
Bring the decision, product inventory, intended transition and the technical assumptions that matter. EmphaSoft can discuss the assessment scope and inputs with your team. Discuss a technical assessment.