Evidence to request
Review live work in your sector or with similar complexity. Ask what the agency was responsible for, what constraints it handled, and what changed after launch. Look for evidence across strategy, UX, content, engineering, and performance rather than visual screenshots alone.
Speak with the people who will lead delivery. A polished sales presentation does not tell you whether the day-to-day team has the experience or authority to solve difficult problems.
| Criterion | Evidence to request | Warning sign |
|---|---|---|
| Relevant experience | Live work with similar users, scale, integrations, languages, or governance | A logo wall with no explanation of responsibility |
| Delivery team | Named leads, roles, seniority, allocation, and location | Only salespeople attend; delivery is described vaguely |
| Arabic capability | Live Arabic journeys, native review process, RTL design and QA examples | Translation is treated as a plugin or final step |
| Technical quality | Architecture rationale, performance process, security practice, QA, and maintainability | Technology is chosen before requirements or explained only with buzzwords |
| Commercial clarity | Comparable scope, assumptions, exclusions, milestones, recurring cost, and change control | A low total with essential work left undefined |
| Support | Warranty, service levels, monitoring, ownership, documentation, and growth options | No clear responsibility after launch |
Define what you need before evaluating agencies
Write a concise brief covering business outcomes, priority audiences and journeys, languages, current problems, content volume, integrations, technical constraints, target date, budget range, internal stakeholders, and expected support. Separate mandatory requirements from preferences.
A clear brief does not prevent agencies from challenging the solution. It gives them enough context to propose a better one and allows you to compare responses against the same problem.
Assess the team, not only the agency brand
Ask who will lead strategy, UX, design, content, Arabic review, engineering, QA, and project management. Meet key people before appointment and understand how much of their time is reserved. Clarify whether work is delivered locally, remotely, through another office, or by subcontractors.
Large teams offer capacity and specialists; small teams can offer continuity and senior attention. Either can work. The risk is a gap between the people who win the pitch and the people who inherit delivery.
Review live work properly
Open recent sites on mobile and desktop. Complete a real journey, switch languages, use forms and search, inspect content quality, and observe performance. Ask which parts the agency delivered and whether the current live site has changed since launch.
Case studies should explain the problem, constraints, approach, responsibilities, and result. Metrics need context: baseline, period, measurement method, and other factors that may have influenced the outcome. A credible agency will distinguish evidence from inference.
Questions that reveal delivery quality
Ask how scope changes are handled, how Arabic is designed and reviewed, how content migration is tested, how accessibility and performance are checked, and what support is available after launch. Clear answers usually indicate a repeatable delivery process.
Present one difficult scenario from your project, such as a delayed integration, conflicting executive feedback, an urgent campaign, or incomplete Arabic content. Ask how the team would respond, what it would need from you, and how schedule or cost impact would be communicated.
- What will you need from our team in the first four weeks?
- Which assumptions could change the price or launch date?
- How do you validate information architecture and user journeys?
- How are Arabic content and RTL components reviewed?
- What is tested before acceptance, and who owns defects after launch?
- How will our team control code, data, design files, domains, and accounts?
Commercial and reference checks
Compare total scope and ownership before price. Confirm payment milestones, VAT, recurring costs, licenses, expenses, warranty, support rates, assumptions, exclusions, and change control. A sustainable commercial model matters because underpriced work often reappears as delays, reduced scope, or change requests.
Speak to two relevant clients if the project is material. Ask whether the named team stayed involved, how the agency handled difficult moments, whether the launch met expectations, how support works, and what the client would do differently.
Make the final decision with a scorecard
Score each finalist against understanding, relevant evidence, team, methodology, Arabic and content capability, technical fit, governance, support, commercial clarity, and cost. Weight the categories before presentations and record reasons for each score.
Use the scorecard to support judgement, not replace it. The best partner should demonstrate that it understands the business, can deliver the difficult parts, communicates tradeoffs honestly, and will leave the organisation with a platform it can operate.
Common questions
Should we choose an agency with experience in our exact sector?
Sector experience is valuable when regulation, users, content, or integrations are specialised. Experience with comparable complexity can be equally important. Balance domain familiarity with evidence of research, adaptability, and avoiding repetitive solutions.
Is a local Riyadh team essential?
Local access can improve workshops, stakeholder engagement, and market understanding, but location alone does not prove delivery quality. Confirm who is available locally, who does the work, and how the wider team collaborates.
How important is the agency's own website?
It is one useful signal for positioning, accessibility, performance, content, and attention to detail, but client evidence and the proposed delivery team matter more. Agencies may prioritise client work over frequent self-redesigns.
What if one proposal is much cheaper?
Normalise the scope. Check content, Arabic, templates, migration, integrations, testing, support, licensing, senior time, and exclusions. The lower price may be efficient, but it may also describe a materially smaller project.