This guide is written for product leaders, founders, and design executives making a real hiring decision. It covers what actually matters when evaluating SF UX firms, what the market looks like in practice, what questions to ask before signing, and what good work at each tier costs.
Start with the brief, not the shortlist
The most common mistake in hiring a UX firm is starting with a list of agencies and working backward to a brief. The better sequence is the reverse: write the brief first, then identify which agencies are suited to it. A brief that's useful for agency evaluation answers four questions:
KEY TAKEAWAY"Improve the UX" is a goal, not a problem. The more specific the problem, the user, and the success metric, the more accurately you can judge whether a firm has solved something like it before.
What is the specific problem?
Not "improve the UX" — that's a goal, not a problem. The problem is where users are dropping off, what workflow is failing, what decision the product is making hard, or what experience is eroding retention. The more specific the problem definition, the more accurately you can evaluate whether an agency has solved something like it before.
Who is the user?
Enterprise software users, medical professionals, crypto-native founders, first-generation smartphone users, and teenagers using a social app are fundamentally different design problems. An agency with ten years of consumer app experience may be genuinely wrong for an enterprise internal tool brief, regardless of how impressive the portfolio looks.
What does success look like?
A number is more useful than a description. "Users complete onboarding" is a goal. "Onboarding completion rate increases from 34% to 60% within 90 days of launch" is a brief. Success metrics shape the entire research and design process — agencies that don't ask for them early are a signal.
What are the constraints?
Timeline, budget, existing design system, technology stack, regulatory environment, stakeholder sign-off requirements, and any decisions already locked. Constraints aren't weaknesses to hide from agencies — they're information that allows the right agency to scope accurately and the wrong agency to self-select out.
Understand what you're actually buying
UX design is not a commodity. Two agencies quoting $80,000 may be proposing fundamentally different things — different research depth, different deliverable contents, different team seniority, different assumptions about how much strategic work the client has already done. Before evaluating proposals, be clear on which of these you're buying:
UX research and strategy
Defining who the user is, what they need, and what design decisions follow. Some agencies do this exceptionally; others treat it as box-checking before wireframes. The difference is visible in how they describe the research phase — firms that do it well can tell you exactly what methods they use, why, and how the output informs design decisions.
UX/UI design
The visual and interaction design of a product — screens, flows, states, components. The quality variation here is large: from interfaces that are visually polished but behaviorally wrong, to ones that are less striking but dramatically more effective to use.
Design systems
The infrastructure that makes UX scale — component libraries, design tokens, Figma files, accessibility documentation, contribution guidelines. In San Francisco, most technology companies expect a design system as part of the deliverable. Make sure the agency has documented experience building and handing off systems engineering teams can actually use.
DesignOps
The operational layer that keeps design consistent as an organization grows — processes, tooling, team structures, and governance. Relevant for companies scaling their in-house design function alongside an agency engagement. Not every engagement needs all four, but knowing which you need shapes which agencies fit.
How to read an SF agency portfolio
The SF market produces exceptionally good-looking portfolios. The polish makes evaluation harder, not easier. Here's how to read past the surface.
Check whether the work is still live
Case study pages are marketing materials. Pull up the live product. Is the work still in use? Has it been significantly revised or replaced? An interface the client quietly replaced within 18 months tells you something the case study won't.
Look for the problem, not just the solution
Strong case studies describe the specific user problem, what the research revealed, what design decisions followed, and what changed as a result. Case studies that only show before-and-after visuals are showing you craft, not capability.
Ask about outcomes
Did the redesign improve conversion? Did the new onboarding reduce drop-off? In San Francisco, where UX is treated as a business function, the best agencies have outcome data. They may not publish it, but they should share it in a pitch conversation.
Check for sector range
A portfolio full of consumer apps from the same era is not evidence of broad capability. Look for work across industries, user types, and product complexity. If the work all looks like the same aesthetic in different categories, ask whether the design follows the brief or the agency's preferred visual language.
Look at the work they don't lead with
Ask for an example of a project that was difficult — where research revealed something unexpected, a direction was scrapped, or the brief changed mid-engagement. How an agency handles complexity is more predictive of a good partnership than how they perform on a clean, well-scoped project.
The SF-specific evaluation criteria
Research methodology transparency
In SF, research is a baseline expectation, not a differentiator. What differentiates agencies is rigor. Ask exactly which methods they use for discovery (interviews, contextual inquiry, diary studies, analytics review, card sorting), when they choose each, and how the output is documented and used. A concrete answer signals a real research practice.
Figma and design system maturity
Ask to see a design system they've built and handed off. Ask whether it includes design tokens (not just components), whether it has accessibility documentation, and whether it was built to a contribution model that let in-house designers extend it after handoff.
Engineering relationship
How does the agency work with development teams? Do they do developer handoff documentation, or hand off Figma files and move on? In SF tech companies the design-to-development relationship is a daily operational reality. Agencies with no opinion on how it should work have not been in those environments.
Accessibility practice
WCAG compliance is increasingly a legal requirement, and SF tech companies treat accessibility as non-negotiable. Ask what level of WCAG compliance the agency designs to by default, how it's tested, and whether the deliverable includes accessibility documentation. Treating accessibility as a late-stage audit adds cost and risk to every project.
AI integration
The SF market has moved quickly on AI tools for research synthesis, rapid prototyping, and design system generation. Ask how the agency uses AI, at which stages, and what guardrails exist for quality control — not because adoption is required, but because no answer means they haven't engaged with the most significant shift in design practice in years.
What to ask before signing
Who will work on the project day-to-day?
Not who presents at the pitch — who is doing the research, making decisions, and showing up to weekly reviews. Senior presentation plus junior execution is the most common source of client disappointment in the agency market.
Can I speak with two past clients directly?
Not a written testimonial — a conversation. Ask references about how the agency handled a difficult moment, how feedback was received and acted on, whether deliverables were usable after handoff, and whether they would hire the agency again.
What does the deliverable package include, exactly?
Ask for a sample deliverable from a comparable past engagement. The difference becomes visible immediately — one will have a comprehensive Figma library with tokens, documentation, and accessibility specs; another will have polished screens in a PDF.
How do you handle scope change?
Design briefs change. How an agency handles that moment — in contract terms, in process, and in attitude — determines whether a mid-project pivot is a problem or a normal part of good design work.
What happens after the engagement ends?
Does the agency offer post-launch support? How are design issues addressed in the weeks after handoff? Agencies with a clear answer have managed enough post-handoff situations to have developed a policy.
What good UX work costs in SF
The SF market commands a premium for several reasons: talent cost, office overhead, and the genuine seniority of the teams that built strong reputations here. That premium is real and worth paying for the right brief — and not worth paying if the brief doesn't require what makes SF firms genuinely better. As a general orientation:
Research and innovation leaders (IDEO, frog) typically engage at $150,000 to $500,000+ for full programs. These numbers are not negotiable through a shortened scope — the process is the product, and compressing it changes what you're buying.
Senior independent studios (Clay, Method, Fantasy) generally start at $50,000 to $200,000 for a defined product UX engagement. The variance reflects scope and team seniority, not quality variation.
Boutique specialists (Neuron, DesignMap, UXReactor) typically fall in the $40,000 to $150,000 range. The ceiling is lower because organizational overhead is lower — an advantage for clients who want direct senior access rather than account management layers.
Early-stage specialists (Mission Control) typically start around $15,000 to $50,000. The model is built for startup constraints without compromising seniority.
One calibration worth making: the cost of a bad UX decision, implemented at scale, is almost always larger than the cost of a good UX engagement. The agencies at the higher end earn their fees by making fewer expensive mistakes and producing work that holds up as the product grows. The question is whether your brief requires that level of rigor — and for most serious product decisions, it does.
A note on timing
UX design is most valuable when it happens before engineering — before the feature is built, before the architecture is locked, before the team has spent six months developing something users won't adopt. The agencies in this directory do their best work when brought in after the product vision is clear but before the implementation details are fixed.
Bringing in a UX firm to redesign a product that has already shipped is more expensive and disruptive than bringing one in to design it right the first time. If your organization is still making the case for UX investment, the timing argument is usually the most persuasive: good UX at the right stage costs a fraction of what bad UX costs to correct.
Ready to build your shortlist?
Browse the 10 reviewed firms and compare them by stage, specialty, and budget.