A client feedback survey is the right tool for one specific job: understanding how a client feels about working with you overall, at the end of a project or on a regular cadence. It is the wrong tool for a much more common job, telling you exactly what's broken on a specific page right now. Confusing the two is why so many agencies send a survey after every project and still get blindsided by complaints that never showed up in a single response.
Below are the questions worth asking, organized by what they're actually good for, plus a template you can copy directly.
What a survey question can and can't tell you
A well-written survey question gets you a pattern: clients consistently rate communication a 4 out of 5 but scope-change handling a 2. That's genuinely useful, it tells you where to invest process improvement. What it can't do is tell you that the pricing table on the client's actual site overlaps at 768px, because nobody is going to type that into a satisfaction survey. For that kind of specific, actionable note, you need feedback pinned to the actual work, which is a different tool for a different moment, covered in more depth in how to collect feedback on a website.
Keep that distinction in mind as you pick questions: a survey measures the relationship, a review round fixes the work.
Know what a "good" score actually looks like before you send
A number with no benchmark is a vanity metric. Before you send anything, decide which metric you're tracking and what target counts as healthy for your category, so a returned score means something instead of just existing. The three worth knowing:
| Metric | What it asks | 2026 benchmark to beat |
|---|---|---|
| CSAT | How satisfied are you with this (1–5 or %) | Cross-industry average sits around 78/100; software/SaaS leads near 80, and top teams target 90%+ |
| NPS | How likely are you to recommend us (0–10) | Anything above 0 is net-positive; above 50 is excellent and above 70 is world-class, though a B2B services NPS near 40 already beats the average |
| CES | How easy was it to work with us | Best paired with the other two rather than tracked alone; a rising CES with a flat NPS usually means friction is building before satisfaction drops |
Two practical notes from the 2026 benchmark data. First, response rate is itself a signal: email surveys typically land at 15–25%, and programs that pair an NPS and a CSAT question see roughly 44% response versus 10–15% for single-metric surveys, so asking two well-chosen questions beats asking one. Second, compare against your own trend, not the global number. A single 78 tells you almost nothing; the same client scoring you 85, then 80, then 74 across three quarters tells you a renewal is at risk before the renewal conversation does.
Project-end feedback questions
Send these once a project wraps, while the experience is still fresh but before the client has moved on to the next thing.
- On a scale of 1 to 5, how satisfied are you with the final result?
- How would you rate our communication throughout the project?
- Did the project stay on the timeline you expected? If not, where did it slip?
- Was the scope of what we delivered clear from the start?
- What's one thing we could have done better?
- Would you recommend us to another business? Why or why not?
- Is there anything you were hoping for that didn't make it into this project?
That last question does more work than it looks like it should. Clients rarely volunteer "I wish you'd asked about X," but they'll answer it directly if you ask, and it often surfaces a scoping gap worth addressing before the next project rather than after.
Ongoing relationship questions (quarterly or per-retainer)
For agencies on retainer or recurring work, a lighter-touch, more frequent check-in catches drift before it becomes a churn risk.
- How satisfied are you with the work this quarter, overall?
- Are our response times meeting your expectations?
- Is there a recurring task or request that's been harder than it should be?
- On a scale of 1 to 10, how likely are you to continue this engagement past the current term?
- Is there anything about the relationship you'd change if you could?
The 1-to-10 "likely to continue" question is a standard Net Promoter Score framing, and it's worth tracking over time rather than treating any single response as a verdict, one bad quarter doesn't mean churn, but three declining scores in a row is a signal worth acting on before the renewal conversation.
A copy-paste form template
| Section | Questions |
|---|---|
| Overall satisfaction | Rate 1–5: final result, communication, timeline adherence |
| Open response | What went well? What could we improve? Anything missing from scope? |
| Forward-looking | Would you work with us again? Would you refer us? |
| Contact | Name, project reference, optional follow-up call opt-in |
Keep the whole form under two minutes to complete. A ten-question survey with five open-text fields gets ignored; a four-section form with mostly ratings and one or two open questions gets answered.
When a survey isn't the right tool
If what you actually need is specific feedback on a specific deliverable, a page, a design comp, a fix that just shipped, a satisfaction survey is the wrong instrument. It asks for a general impression when what you need is a precise note tied to a specific element. "How satisfied are you with the homepage redesign, 1 to 5" tells you almost nothing actionable. "The hero button color is too close to the background on mobile" is something a developer or a coding agent can act on immediately.
This is the gap most agencies hit without noticing: they run a great end-of-project survey and still spend the actual project drowning in vague email threads because there was never a structured way to leave feedback on the work itself, only on the relationship after the fact. If you're building out a survey process, it's worth pairing it with a real review workflow for the in-project feedback; our client approval process breakdown covers exactly that half of the loop.
Turning recurring questions into a public FAQ
If the same three or four questions keep coming up across projects, "how do revisions work," "what's included in this retainer," it's worth publishing the answers as an actual FAQ page rather than re-answering them in every kickoff call. A short FAQ section marked up with FAQ schema also has a shot at showing directly in search results, which cuts down on the "quick question" emails before a project even starts. Our schema markup generator turns a plain list of question-and-answer pairs into valid, paste-ready JSON-LD without needing to hand-write the structured data.
When to actually send it
Timing changes the response rate more than the questions do. Send a project-end survey within a day or two of delivery, while the work is still fresh, not weeks later when the client has moved on to the next priority and answers it on autopilot. For ongoing retainer relationships, quarterly is usually the right cadence, monthly tends to produce fatigue and declining response rates, while yearly is too infrequent to catch a relationship drifting before it's already a renewal problem.
Avoid sending a survey in the same email as an invoice or a request for something else. Bundling a satisfaction question with a payment reminder skews responses toward whatever emotion the invoice triggers, not an honest read on the work itself. A short, standalone message, ideally from a person rather than an automated no-reply address, gets both better response rates and more candid answers.
What to do with the answers
A survey nobody reads is worse than no survey, it signals to clients that their time didn't matter. Build a five-minute habit around it: after every response, someone on the team reads it, and any score below a 3 gets a follow-up, not an automated thank-you email. Track the ratings over time by project type or by client manager, patterns that don't show up in any single response often show up clearly across ten of them.
Where Shotline fits
Surveys tell you how a client feels about working with you. For the moment-to-moment work itself, feedback that's specific enough to act on needs to live where the work lives. Shotline lets clients comment directly on the live page with a pin tied to the exact element, no account required, so "the button color is off" comes with the selector and screenshot attached instead of a vague note in a satisfaction score. For the full comparison of what to look for in a client feedback tool beyond surveys, see the best client feedback tools for agencies.
Shotline is free to try for 14 days, no card required, then from $19/mo (billed annually; $25 month-to-month) with unlimited seats. Start a free trial or see how it works on the Shotline homepage.
