Pulse
Scheduled Sep 27, 2026Surveys that do not bill you per response.
Replaces
- SurveyMonkey · Momentive Global Inc. — $30 per seat per month (Team Advantage), about $144,000/yr at 400 seats
- Typeform · Typeform S.L. — $83 per org per month (Business), about $996/yr at 400 seats
- Qualtrics CoreXM · Qualtrics International Inc. — no published list price
Rented, per year
$144,000
SurveyMonkey list price at 400 seats.
Owned, per year
$4,920
AWS infrastructure only. Azure is $6,240.
Build
9 weeks
2 engineers. To parity on the core features above, including migration of the question library, active surveys and historic response data from SurveyMonkey, Typeform or Qualtrics exports.
Difficulty
3/5
Well-understood shape. Low delivery risk.
Reference organisation: 60 staff seats, 40 surveys a year, 1.2M responses a year with launch spikes of ~400 writes a second, 5 years of retained responses, single region, prod + non-prod. 3-year Compute Savings Plan on Fargate / 3-year reservation on Container Apps. Aurora minimum 0.5 ACU. DynamoDB and Cosmos DB on demand. Invitation email delivery billed through Dispatch and not included. Excludes LLM gateway token cost for open-text coding. Figures are infrastructure only and exclude the build.
What it is
Pulse is a survey and form platform: a question library, branching and display logic, quotas, multi-channel distribution, and analysis with cross-tabs and open-text coding. It is aimed at organisations running research, customer feedback and operational forms at a scale where response caps have started shaping the research design. Pulse removes the two constraints that distort survey programmes: a per-response ceiling and a per-seat licence for the analysts who read the results.
Why this one stops making sense
- Typeform Business lists at $83 a month and stops at 10,000 responses. A single customer survey to a 200,000-person list makes the plan irrelevant on the day you send it.
- SurveyMonkey Team Advantage is $30 per user per month with a three-user minimum, so every analyst who wants to look at the data is another licence.
- Qualtrics publishes nothing. Its pricing page carries no figures at all, which makes internal budgeting for a research programme an exercise in asking the vendor.
- Response caps change research design. Teams sample smaller, drop follow-ups and shorten fielding windows to stay inside a tier, and then treat the result as if the constraint had been methodological.
- Raw response rows — including partials, timings and drop-off points — are the difference between analysis and a summary chart, and they are the part that arrives last and least completely on export.
What ships
- Question library with validated blocks, matrix, ranking, slider, NPS and open text
- Display logic, skip logic, randomisation, piping and carry-forward across blocks
- Quotas and screening with automatic termination and quota-full handling
- Distribution by email, unique link, anonymous link, QR code and embedded iframe
- Partial response capture, including drop-off point and per-question dwell time
- Cross-tabulation and significance testing with weighting by known population margins
- Open-text coding with an assisted first pass and human review of every code
- Multi-language surveys with per-language logic and translation status tracking
- Anonymous response mode with a minimum-count threshold on any breakdown
- Direct SQL access to raw response rows and an event stream into your warehouse
Data model
Stack
- frontend
- Next.js 15 App Router for authoring and analysis; a separate lightweight respondent bundle
- backend
- NestJS on Node 22 for authoring and analysis; a minimal intake service for responses
- database
- PostgreSQL 16 with answers in a partitioned table, partitioned by survey and month
- cache
- Redis 7 for quota counters, in-flight response state and rate limiting
- queue
- BullMQ for invitation dispatch, response materialisation and export
- auth
- OIDC for staff via the customer IdP; signed single-use tokens for respondents
- search
- PostgreSQL tsvector over open-text answers and survey metadata
- ai
- Optional open-text coding and theme clustering through an LLM gateway, with human review
Parity, honestly
What we match, and what we do not.
The rows marked No are the important ones. If one of them is the reason you bought SurveyMonkey, keep buying it — and we will tell you that on the first call rather than the fifth month.
| Capability | SurveyMonkey | Ours | Note |
|---|---|---|---|
| Branching, display and skip logic | Yes | Yes | — |
| Quotas, screening and termination | Yes | Yes | — |
| Multi-channel distribution | Yes | Yes | Email distribution runs through Dispatch, so a 200,000-person send is delivery cost, not a plan tier. |
| Unlimited responses | No | Yes | A response costs a row and a few bytes. It should not cost a plan upgrade. |
| Unlimited free analyst and viewer seats | No | Yes | — |
| Cross-tabs, weighting and significance testing | Yes | Yes | Weighting by known population margins, with the weights exposed rather than hidden. |
| Raw response rows including partials and timings | Partial | Yes | In your own database, queryable by your own analysts. |
| Respondent panel and purchased sample | Yes | No | Qualtrics and SurveyMonkey both sell access to millions of vetted respondents. That is a marketplace with real recruitment and quality operations behind it. We cannot build it and you should keep buying it if you need sample. |
| Advanced research methodology: conjoint, MaxDiff, TURF | Yes | No | Qualtrics CoreXM ships validated conjoint and MaxDiff designs with the analysis to match. Building those correctly is a statistics project, not a feature, and getting them subtly wrong is worse than not having them. |
| Typeform's conversational form experience and template gallery | Yes | Partial | We ship one-question-at-a-time forms in your design system. Typeform's polish and its gallery of hundreds of templates are a genuine product advantage. |
| Vendor-run SOC 2, ISO 27001, HIPAA and FedRAMP | Yes | No | Qualtrics holds authorisations that matter for government and health research. Pulse inherits your cloud posture only. |
| Anonymous response guarantees | Yes | Yes | Enforced by a minimum-count threshold applied in the query layer, in code you can read. |
| Open-text coding and theme analysis | Yes | Yes | Assisted through your LLM gateway with every code reviewable and editable. |
| Offline and kiosk data collection | Yes | Partial | PWA with offline queueing. Dedicated kiosk hardware management is out of scope. |
Reference architecture
Built twice, on purpose.
Every application in Techtons ships with a production architecture for AWS and one for Azure, using the same diagram grammar so you can read them side by side. Your cloud, your account, your bill.
Amazon Web Services
- Compute
- API Gateway HTTP API plus Lambda for response intake; ECS Fargate for authoring, analysis and workers
- Data
- Aurora Serverless v2 PostgreSQL; DynamoDB holding published survey definitions and quota counters
- Storage and edge
- S3 for uploaded files, exports and the compiled respondent bundle, CloudFront in front
- Identity
- Cognito federated to the customer IdP for staff; respondents are unauthenticated or token-bearing
- Observability
- CloudWatch and OpenTelemetry; intake latency and quota-counter drift both alarmed
- Infrastructure as code
- Terraform, one module per environment
Services
The decision worth arguing about
The response path never touches the application database, and the reason is the traffic shape rather than security. A survey link goes out to 200,000 people at 09:00 and roughly a fifth of the responses arrive in the first twenty minutes, which is a load profile nothing else in this catalogue has: near-zero for days, then several hundred writes a second, then near-zero again. So intake is API Gateway to Lambda, reading the published survey definition and quota counters from DynamoDB and writing each response to an SQS FIFO queue keyed by response identifier; Fargate workers drain it into Aurora at whatever rate Aurora is comfortable with. Aurora Serverless v2 can scale for the spike, but it scales in seconds while the spike arrives in milliseconds, and a survey that drops responses on launch morning is not a survey you can rerun. The trade-off is that one logical entity now lives in two stores. A published survey definition is copied into DynamoDB, and edits take up to sixty seconds to propagate, so Pulse locks structural edits while a distribution is live and offers a clone-and-branch instead — which is also the methodologically correct behaviour, since changing a question mid-field quietly ruins the dataset. Quota counters live in DynamoDB with conditional writes rather than in Postgres, and are reconciled against the authoritative row count after the queue drains.
Microsoft Azure
- Compute
- Container Apps for authoring, analysis and workers; a dedicated intake container app with a pinned minimum replica
- Data
- PostgreSQL Flexible Server; Cosmos DB for published definitions and quota counters
- Storage and edge
- Blob Storage for uploads and exports, Front Door Standard for edge and WAF
- Identity
- Entra ID for staff; signed single-use tokens for respondents
- Observability
- Azure Monitor and Application Insights, with cold-start count as an explicit metric
- Infrastructure as code
- Bicep, azd-compatible layout
Services
The decision worth arguing about
Container Apps scaling to zero is the feature everyone reaches for on Azure and it is exactly wrong for the intake service. A scaled-to-zero replica takes a few seconds to come back on the first request, and the first request here is a respondent clicking a link in an email — the moment where a blank few seconds costs you the response and, at the top of the funnel, some percentage of the whole sample. So the intake app is pinned at a minimum of one replica permanently, with the scale rule tuned to add replicas on concurrent requests rather than on queue depth, and only the analysis and export workers are allowed to scale to zero. That pin is a fixed floor of roughly $35 to $50 a month per environment even in a month where nobody fields a survey, which is the price of not losing respondents to a cold start. AWS avoids the question entirely because Lambda's cold start is tens of milliseconds for a small handler and API Gateway is billed purely per request, and that is the honest reason the AWS figure below is lower for a workload that is idle most of the time.
Sources
Every price on this page, with the page we read it from and the date we read it.
| Product | Plan | List price | Checked | Source |
|---|---|---|---|---|
| SurveyMonkey | Team Advantage | $30 / seat/month | 2026-09-05 | www.surveymonkey.com/pricing/ Billed annually, starting at three users. Team Premier lists at $92 per user per month; Enterprise is quoted by sales. |
| Typeform | Business | $83 / org/month | 2026-09-05 | www.typeform.com/pricing/ Billed annually and capped at 10,000 responses a month. Plus lists at $50 a month for 1,000 responses; Basic at $25 for 100. |
| Qualtrics CoreXM | CoreXM | Not published | 2026-09-05 | www.qualtrics.com/pricing/ Qualtrics publishes no list price. Its pricing page shows no dollar figures and states that pricing and plans differ by industry security requirements and product availability, routing to a sales conversation. |
Should you still be paying for SurveyMonkey?
Two weeks, fixed price. We audit what you actually use, map it against Pulse, cost the replacement on your own AWS or Azure account, and give you a delivery plan. If the honest answer is to keep the licence, that is what the report will say.