A UK CEO we spoke to last month was talking to 5 development agencies about a £220k build. All five had polished pitch decks. All five claimed relevant experience. All five gave similar timelines. He could not tell them apart.
That is the partner selection conversation across UK and US businesses in 2026. Buyers who use a structured checklist pick well. Buyers who go on chemistry and pitch deck quality often pick wrong and pay the cost 6-12 months later.
This article is a candid 12-point checklist for CEOs, COOs, and CTOs evaluating software development partners. What to ask. Red flags and green flags. What to ignore. Real cost of the wrong choice. Written from the buyer's side, though we are ourselves a development partner.
The 12-Point Partner Fit Checklist
Twelve specific checks that surface real capability.
1. Show me 3 shipped products in your domain. Not "similar to your domain." Actual products in your specific vertical or use case. If they cannot show three, they lack domain depth.
2. Let me talk to 3 clients from 12+ months ago. Reference clients under 6 months old have not been through the delivery cycle where problems surface. 12+ month references tell you what the agency looks like when things get hard.
3. Introduce me to the team who will actually work on this. Not just sales and account manager. The engineers, tech lead, and PM who will be on your project. If they cannot arrange this before contract, walk.
4. Explain your SDLC and communication cadence in specific terms. Not "agile" as a word. Specifics: sprint length, standup frequency, demo cadence, retrospective process, tools used, artifacts produced. Agencies without specific process usually improvise.
5. Show me a real PM artifact from another client (redacted). Sprint plan, delivery roadmap, weekly status report, retrospective doc. Real artifacts prove real process; refusal to share signals process does not exist.
6. Explain your contract termination clause and notice period. Termination for convenience should be 30-60 days with fair payment for work in progress. Longer than 90 days or restrictive clauses are red flags.
7. Clarify IP ownership, licensing, and open source policy. Full IP transfer on payment should be default. Open source components disclosed. License compatibility documented. Agencies vague on this create legal risk downstream.
8. Verify data residency and security posture. SOC 2 Type II, ISO 27001, or equivalent. Clear data residency (UK, EU, US). Documented security practices. Especially critical for regulated industries.
9. Explain your rate card and total cost structure transparently. Headline day rate matters less than all-in cost including overhead, PM, infrastructure, change control. Ask for typical total cost on comparable projects.
10. Tell me what you would say NO to. Agencies that say yes to everything are a red flag; they lack judgement or plan to over-scope. Agencies that describe what they would decline demonstrate real judgement.
11. Describe your bench depth and staffing plan. How many engineers total, what happens if my assigned engineer leaves, coverage during holidays and illness. Small agencies with 3 engineers cannot cover the same continuity as 20-engineer agencies.
12. Walk me through your escalation path when things go wrong. Communication protocol when delivery slips, formal escalation to leadership, how disputes get resolved. Agencies without a defined escalation path escalate through chaos.
Agencies that pass 10+ of these 12 are candidates worth serious consideration. Agencies that fail 4+ are best walked from. The Gartner IT services research publishes ongoing analysis of enterprise IT vendor evaluation.
Red Flags vs Green Flags
Red flags to walk from:
No reference clients from past 12 months (nothing has been through delivery)
Salespeople dominate every call; engineers never appear
Generic case studies presented as domain expertise
Refusal to name specific engineers before contract signing
Non-negotiable standard contract terms
Vague on IP ownership or security posture
Rate card obscured behind bespoke pricing dance
Say yes to every requirement without trade-off discussion
Green flags to trust:
Put engineers in the room during evaluation
Quote realistic timelines even when a shorter one would win the deal
Tell you what they will not do (technology fit, scope discipline)
Transparent about pricing structure and all-in cost
Comfortable being cross-examined on past work details
Willingly share redacted PM artifacts
Push back constructively on your assumptions
The GOV.UK procurement policy notes publish public sector supplier evaluation guidance that overlaps with private sector best practice.
What NOT to Get Distracted By
Five distractions that pull buyer focus from what matters.
Fancy pitch deck. Beautiful decks are cheap to produce. Correlate weakly with delivery capability. Focus on underlying content, not design.
Case studies from unrelated industries. Agency showing 50 case studies across every vertical rarely has deep expertise in yours. Three deep case studies in your vertical beat 50 shallow ones.
Buzzword compliance. AI, blockchain, Web3, machine learning, quantum-ready, agentic sprinkled across every capability. Real depth has specific examples and honest limits; buzzword-heavy pitches without specifics signal shallow capability.
Big-name logos where they were a tiny subcontractor. Working on a piece of a Fortune 500 project as one of 15 vendors is not "delivering for Fortune 500." Ask what specifically they did, at what scope, with what team.
Cheap headline rate that hides expensive total. £400/day headline that becomes £850/day all-in after adding project management, testing, DevOps, infrastructure, and change control. Ask for total cost on comparable projects.
Real Cost of the Wrong Partner Choice
Direct cost. 30 to 60 percent budget overrun on the original scope. Rework, rebuild, or migration to a second agency. On a £150k project, additional cost typically £50k to £120k.
Delay cost. 3 to 9 months of delay from what was planned. Missed market windows, delayed revenue, lost competitive position. Often the largest hidden cost.
Opportunity cost. Founder time and executive attention spent managing a struggling relationship instead of building the business.
Recovery cost. Second agency often needs 2-4 months to understand what the first agency built before making changes. Domain knowledge lost in transition.
Total real cost of a wrong partner choice on a £150k project typically £120k to £280k additional across all four categories, plus 6-9 months of lost business value.
What We See When Founders Compare Us
We are ourselves a software development partner. Across founder evaluations at WhiteStone, three patterns recur.
Founders who use structured evaluation pick better regardless of who they choose. Founders who apply a checklist (whether ours or their own) end up with more successful engagements even when they choose a different agency. Structure beats gut feel.
Reference calls from 2-3 year old clients reveal the most. Recent clients speak to enthusiasm. 2-3 year old clients speak to delivery, follow-through, and post-launch support. We proactively offer references from our earliest UK property developer engagement because that history matters.
Founders who ask what we would decline get the best fit. When we tell buyers we would decline a specific engagement (wrong tech fit, unrealistic timeline, scope beyond our capability), some walk. Those who stay tend to become durable clients because we picked the right work together.
You can see our shipped work at our portfolio. If you want to run us through your 12-point checklist, book a partner-fit call with WhiteStone.
Common Failure Modes
Three failure modes we see repeatedly.
Choosing on headline rate alone. Cheap headline becomes expensive total after PM, testing, DevOps, infrastructure, change control. Focus on all-in cost.
Skipping reference calls to close the deal faster. Reference calls take 1-2 hours. They save 6+ months of pain. False economy to skip.
Ignoring the "what would you say no to" answer. Agencies that say yes to everything become expensive later through scope creep or over-promise. Judgement to decline builds durable partnerships.
Frequently Asked Questions
What questions should I ask a software development agency?
Twelve core questions covering domain shipped products, 12+ month reference clients, engineers on your project, specific SDLC, PM artifacts, contract termination, IP ownership, data residency, all-in pricing, what they say no to, bench depth, and escalation path. Agencies that pass 10 of 12 are candidates; agencies failing 4+ are best walked from.
How do I check if an agency really built the projects they claim?
Ask for URLs to live products they built (not archived screenshots). Talk to the client contact who commissioned each project. Ask what the agency delivered vs the client team or other vendors. Big projects involve many parties; be sure the agency's contribution matches their claim.
What are the biggest red flags in software agency selection?
No reference clients past 12 months, salespeople dominate every call, refusal to name engineers before contract, non-negotiable contract terms, vague on IP or security, cheap headline rate hiding all-in cost, saying yes to every requirement. Any 3+ together should trigger walking away.
What is a realistic cost for a custom software build in 2026?
Highly variable by scope. Rough bands: £30k-£120k MVP over 8-16 weeks, £150k-£400k production platform over 6-12 months, £500k-£2m enterprise programme over 12-24 months. Ask agencies for total cost on comparable projects, not day rates.
How do I evaluate a software development agency's team quality?
Meet the engineers who will actually work on your project before signing. Review their LinkedIn profiles and GitHub or portfolio work. Ask about technical background, years of experience, past project involvement. Agencies that hide their engineers until after contract are hiding something.
The One Thing to Remember
Choosing a software development partner is the single most consequential decision of the build. Get it right and the project ships on time at budget. Get it wrong and it overruns 30-60 percent, slips 3-9 months, often needs a second agency to complete. Twelve specific checks separate real capability from good pitch decks.
Buyers who use structured evaluation pick well regardless of vertical or budget. Buyers who go on gut feel and pitch deck quality pay for years.
If you want a candid conversation about partner fit, browse our custom software development services or come to the call.
