The short answer
What matters most
Choose a software development partner by testing relevant project experience, the proposed team, delivery responsibilities, and ownership arrangements against the same project brief. Use a focused assessment when a major technical dependency is still unknown. A low quote or a recognizable client logo alone does not establish that the partner can deliver your workflow.
- Compare proposals against one brief and the same acceptance criteria.
- Ask for evidence of the work the proposed team will actually own.
- Agree on communication, account access, and handover before starting.
Compare before you commit
Decision at a glance
| Selection area | Evidence to request | A reason to investigate further |
|---|---|---|
| Relevant experience | A walkthrough of comparable workflows and the partner’s exact contribution | A list of logos with no explanation of delivered scope |
| Team availability | Named roles, allocation, technical leadership, and replacement process | The proposed team remains undefined when the engagement starts |
| Delivery control | Example milestones, acceptance checks, risk reporting, and change process | Dates are promised before access and dependencies are understood |
| Ownership and continuity | Repository access, account ownership, documentation, and exit deliverables | Source code or deployment knowledge depends on one person |
What should you decide before approaching a partner?
You need a business problem and someone who can discuss it—not a complete technical specification. If scope is clear, compare delivery proposals. If an integration or existing system is poorly understood, define a focused assessment first. If you only need extra capacity, describe the role and the leadership already available.
Discuss your project with SaltStart with the business need
Write down who will use the software, what problem it should solve, and what would count as a useful first release. Separate essential capabilities from later ideas. A specific brief makes proposals easier to compare.
Use a single comparison brief for every shortlisted partner. Include the main workflow, expected users, required integrations, data sensitivity, budget constraints, and the decisions you have not made. If one proposal assumes an existing design and another includes design work, their totals are not comparable. Ask each partner to flag missing information instead of filling the gaps silently.
Ask for relevant evidence
Ask to see a project with similar users, integrations, or operational constraints. Discuss what the vendor actually built, who worked on it, what went wrong, and how the system is supported today. A recognizable client logo is not a substitute for a relevant project story.
Walk through one case in detail. Ask which capabilities shipped, which were planned, and what evidence supports the stated outcome. A working demonstration, an architecture explanation, or an authorized customer reference can answer different questions. Confidentiality may limit what a vendor can share; it should still be possible to explain responsibilities and technical decisions without disclosing a client's sensitive information.
Meet the people doing the work
Understand who handles development, technical decisions, testing, and day-to-day communication. Ask about availability, replacement arrangements, timezone overlap, and whether proposed specialists are assigned full-time or shared.
Ask who reviews pull requests, who approves releases, and who makes a decision when a requirement is ambiguous. Request a sample weekly update that shows completed work, risks, and decisions needed from you. Confirm how holidays, absence, and staff changes are handled. Access to a strong salesperson does not establish the availability of the proposed delivery team.
Compare the delivery plan
Look for assumptions, milestones, acceptance criteria, dependencies, and a process for changing scope. A detailed plan is useful only if the team can explain how it will adapt when new information appears.
Choose one important user journey and ask how the partner would deliver and validate it. For an order approval tool, discuss incorrect submissions, permissions, duplicate requests, and failed notifications as well as the happy path. You are looking for a team that identifies dependencies and can explain tradeoffs in language your stakeholders understand.
Clarify ownership and handover
Agree on access to source code, cloud accounts, design files, documentation, and deployment processes. Record intellectual-property ownership, support responsibilities, and exit arrangements in the contract before work starts.
Distinguish ownership of custom work from licenses for pre-existing components. Ask whose accounts hold the code, hosting, analytics, and external subscriptions, and when you receive access. Agree on what happens if the engagement ends before completion: documentation, unfinished work, credentials, data exports, and transition assistance should all have an owner.
Use the first engagement to reduce uncertainty
An assessment or focused pilot can reveal technical risks and show how the team communicates. Agree in advance on what the pilot must demonstrate and what will be delivered at the end.
Keep the pilot representative of the real risk. A polished screen does not validate a difficult integration; a technical prototype does not validate user adoption. Define the question, the evidence required, the deliverables you retain, and the next decision. At the end, compare observed communication and delivery against the promises made during selection.
Illustrative example · not a client result
Comparing partners for a customer portal
A US distributor wants customers to see order status from its ERP. One vendor proposes a fixed project; another proposes a managed team.
- Give both vendors the same order-status workflow, permission rules, and ERP documentation.
- Ask each to demonstrate how failed data refreshes and account isolation will be tested.
- Separate included work, client dependencies, and optional features in their proposals.
The decision
If ERP access is unproven, commission a bounded integration assessment before committing to the full portal. Select the next step using the evidence from that assessment.
Planning with a US team
For a US–India engagement, specify overlap in your actual US time zone, including daylight-saving changes. Agree who handles decisions outside shared hours. Identify the contracting entity and have your procurement and legal teams review the proposed ownership and data-access terms.
Use this in your next discussion
Your planning checklist
Work through these points with your stakeholders or shortlisted partner. Record the owner of any unanswered question.
- Define the first complete user journey and its business owner.
- List integrations, access dependencies, and data restrictions.
- Request a relevant project walkthrough and clarify the vendor’s role.
- Meet the proposed delivery lead and agree on communication overlap.
- Record acceptance, ownership, support, and transition responsibilities.
Common questions
Should a US company work with an offshore development partner?
It can be a fit when the partner’s capability and working arrangements match the project. Evaluate shared working hours, response expectations, access controls, delivery ownership, and evidence of similar work. An office address alone does not establish delivery suitability.
Is a paid discovery phase always necessary?
No. A well-defined change may be estimated directly. Discovery is useful when uncertainty about workflows, integrations, or existing code could materially change scope. Ask for defined outputs and the decision that discovery will support.
How many proposals should we compare?
Compare enough qualified candidates to understand the trade-offs, using one brief. A larger shortlist is not useful if you cannot examine the proposed teams and assumptions. Resolve major scope differences before comparing totals.
What should I share with Salt to start this discussion?
A short description is enough: your intended users and the problem they need solved; existing software and integrations the project must work with; your target date and the decisions you still need help making. You can leave confidential documents out of the initial enquiry. We will clarify the requirements and agree whether a project, team, or assessment is the appropriate next step.
Explore Salt’s work
Relevant project experience
MyPowerStock demonstrates custom inventory and sales application work. The American digital engineering partnership demonstrates engineering support within an existing team. Use them to ask about the kind of delivery you need, rather than treating a client logo as proof of fit.
MyPowerStock: Custom Inventory & Sales Enablement Portal
Salt developed MyPowerStock for Power Promotions: a custom inventory portal and sales enablement platform, with product engineering and CI/CD work included in the engagement.
Read the case study ↗Digital engineeringDevelopment Support for a US Digital Engineering Company
Salt supported an American digital engineering company with additional engineering capacity across requirements, development, testing, and deployment.
Read the case study ↗Your next step
Shortlisting a partner for a real project?
Salt builds custom applications and supports existing engineering teams. Start with the business workflow you need to improve; we can discuss the delivery scope, the people it needs, and the questions to resolve before estimating.
A useful starting point
- Your intended users and the problem they need solved
- Existing software and integrations the project must work with
- Your target date and the decisions you still need help making
A few sentences are enough. You do not need a finished specification or confidential documents to start.
Discuss your project with SaltRead client reviews