Two teams, two scores
Support reports CSAT from a filtered view, the CX team from all responses, and finance from last quarter's export. Each number is right for its own filter, and nobody wrote down which filter is official.
SurveyMonkey · Reporting & Analytics
SurveyMonkey data and analytics services that turn responses into decisions
SurveyMonkey data and analytics is the work of turning survey responses into numbers people trust and act on: defined metrics, surveys built to be segmented, analysis in SurveyMonkey, and reporting in Power BI, Tableau or Salesforce. Twopir Consulting designs the metrics and the reporting model, sets up the analysis, and connects the data to the tools your teams already use. The goal: one definition per metric, one source per number, and a named owner for every decision it feeds.
12+
500+
250+
40+
Trusted by 500+ organizations — growing and mid-market companies and enterprise teams that rely on Twopir Consulting for CRM data, reporting and integration work.
Where We Report on Survey Data
- Salesforce Partner
- SurveyMonkey Analyze
- Salesforce Reports
- CRM Analytics
- Tableau
- Power BI
- SPSS
Plenty of responses, not enough answers
Most organizations do not lack survey data. They lack a number that two departments agree on, and a way to see it by the segments that matter. These are the six reporting problems we are asked to fix most often.
Results that cannot be segmented
The question everyone asks is "by region, by product, by account tier". The survey never captured those keys, so the only cut available is the one respondents typed in themselves.
Reporting by monthly export
Someone downloads a spreadsheet, rebuilds the same charts and emails a deck. SurveyMonkey keeps an export downloadable for 14 days, so last cycle's working file is often the only record of how a number was made.
Open-text answers nobody reads
Comments hold the reasons behind a score, yet they sit unread because reading hundreds a month is nobody's job. The score gets reported; the explanation does not.
Responses in Salesforce, but no reports
The integration syncs responses, but reporting was never switched on for the response objects, or each answer sits on its own child record and nobody built a report type that reads them together.
A BI connection on one person's laptop
The Power BI report was built in one analyst's Power BI Desktop with their own access token. It works until they leave, change machines or rebuild the survey the report depends on.
Definition
SurveyMonkey analytics, defined
SurveyMonkey data and analytics services turn survey responses into trustworthy reporting: they define the metrics a survey must produce, design the survey so its data can be segmented, analyse the results in SurveyMonkey, and deliver them to the place each audience makes decisions, whether that is SurveyMonkey itself, a BI tool such as Power BI or Tableau, or Salesforce reports. The work covers both sides of a number: how it is calculated, and who sees it, how often, and what they are expected to do about it.
Three parties are involved. SurveyMonkey collects the responses and provides the analysis tools: rules and saved views, crosstabs, exports, text and AI analysis, and connectors to Power BI and Tableau. Many of these depend on the plan, as the SurveyMonkey Analyze help center explains. Your reporting tools (Salesforce, Power BI, Tableau, a warehouse) own the dashboards, sharing and refresh on their side. Twopir Consulting defines the metrics, designs the reporting model, configures the analysis and builds the reports, and documents which number comes from where.
When do you need a partner? A single survey read by the person who built it does not need one. It becomes worth it when several teams report on the same score, when survey data has to sit beside CRM data, when leadership wants a trend across surveys and years, or when open-text volume is beyond what anyone can read. We help growing and mid-market companies solve complex CRM, integration and business system challenges, and we bring the same reporting discipline to enterprise organizations running survey programmes across many teams.
Designing a whole survey programme from scratch? That is our SurveyMonkey consulting and implementation services. Deciding which systems survey data should reach is covered under SurveyMonkey integration services. This page is about what happens to the data once you need answers from it.
Where each reporting question is best answered
There is no single right place to report on survey data. Each path suits a different audience and a different question, and each has documented limits. We pick per audience, and we keep the number of places a metric is calculated as small as possible.
| Path | Best for | Plan | Limits that shape the design |
|---|---|---|---|
| SurveyMonkey Analyze | Survey owners and team leads exploring one survey: filtered and compared results, saved views, crosstabs, text and AI analysis. | Varies by feature: crosstabs are listed for Team Premier and above, and AI analysis for Advantage, Premier, Team and Enterprise plans. | Survey-by-survey by nature. Crosstabs allow 5 column and 10 row questions and are not available above 100 questions. |
| Exports | Board packs (PDF, PowerPoint), analyst work in Excel, and research teams in SPSS. | Paid feature. | A snapshot, not a feed. Files stay downloadable for 14 days, and above 100,000 responses (or respondents times questions) only CSV is offered. |
| Power BI or Tableau | Trends across surveys and periods, survey data next to finance or operations data, and audiences who already live in a BI tool. | Power BI: Enterprise add-on, Power Users only. Tableau: Enterprise. | Power BI connects through Power BI Desktop on Windows with an access token; scheduled refresh runs once published to PowerBI.com. Tableau receives Extracts. |
| Salesforce reports and dashboards | Account teams, support leads and sales leaders who need scores beside pipeline, cases and renewals. | Needs SurveyMonkey for Salesforce, an Enterprise add-on. | Only responses collected after a mapping exists sync by default. The reporting model depends on the mapping choice, covered below. |
| SurveyMonkey API | Loading full response data into a warehouse on your own schedule, when the packaged connectors stop short. | Paid plan with API access for detailed response data. | An engineering project with rate limits of its own. How we build API extractions. |
Getting more from Analyze before you reach for BI
Most survey questions can be answered inside SurveyMonkey once the views are set up properly. We configure the analysis your teams return to every cycle, so nobody rebuilds the same filter by hand, and we document the limits that decide when a question has to move to a BI tool.
Rules and Saved Views
Filter rules narrow results to a subset, compare rules put answer options from one question side by side, and show rules pick the pages or questions on screen. Combinations are saved as views.
- What we set up
- One named view per recurring audience, so "the support view" means the same filter for everyone
- Watch for
- A crosstab built on a saved view does not update when the view is edited; it has to be recreated
Crosstab Reports
Several questions compared at once in a table, each cell showing count, percent of total, row percentage and column percentage. Crosstabs can be published by link, keep updating as responses arrive, and download as CSV.
- Plan
- Paid; listed for Team Premier and above
- Watch for
- Up to 5 column and 10 row questions; Multiple Choice, Checkboxes, Dropdown and Image Choice only; no combined filters, and no saved views that contain compare or show rules
Significance and Multi-Survey Analysis
Statistical significance shows whether a difference between groups is likely to be real, and multi-survey analysis reads several surveys together, such as the same pulse run each quarter.
- Plan
- Listed for Team Advantage and above
- Watch for
- Both only work if the questions and answer scales stay identical between surveys, which is a design rule, not a setting
Charts and Question Summaries
Question summaries give a per-question chart and data table, and export as presentation-ready PDF or PowerPoint slides, or as XLS and CSV with counts and percentages already calculated.
- What we set up
- A standard export recipe per report, with the view it uses named in the file
- Watch for
- A summary export is a snapshot; nothing in the deck updates when new responses arrive
Individual Response Exports
One row per respondent in XLS or CSV for your own analysis, PDF for reading single responses, or SPSS (a .SAV file) for research teams. Custom variables are included in these exports.
- Plan
- Paid feature; SPSS 17.0.1 or later recommended
- Watch for
- Above 100,000 responses, or respondents times questions, only CSV is available; exports are kept for 14 days
Benchmarks
Questions taken from SurveyMonkey's Question Bank can be compared with results from other customers who asked the same question, which gives a score some outside context.
- What we set up
- Question Bank wording where a benchmark matters, decided before the survey goes live
- Watch for
- Benchmarks cover Question Bank questions only, so the wording decision comes before launch, not after
Reading open text at scale, with the controls on
SurveyMonkey's AI analysis features read comments and summarise results far faster than a person can. They are also plan-gated, and some send survey content to outside model providers. We decide which ones to switch on, for which surveys, and how their output is checked before it reaches a report.
| Feature | What it does | How the data is processed |
|---|---|---|
| Analyze with AI | Answers questions about the results in a chat, returning insights, charts and summaries. | Sends the survey title, questions and responses to OpenAI. |
| Thematic Analysis | Groups open-text answers into themes. | Uses a self-hosted Anthropic model and OpenAI, and masks common sensitive data types. |
| Sentiment Analysis | Classifies open-text answers as positive, negative or neutral, and the result can be used as a filter. | SurveyMonkey's own machine-learning models. |
| Response Quality | Detects low-quality responses (SurveyMonkey's response quality detection), so they can be reviewed before they are counted. | SurveyMonkey's own machine-learning models. |
| Custom tags | Several tags per text response, usable as filters. On Enterprise, tags can be applied automatically as responses arrive. | Your own tag list; the rules are yours to maintain. |
Survey scores in Salesforce, reported beside the pipeline
With SurveyMonkey for Salesforce, responses become Salesforce data and can be reported with native reports and dashboards. How easy that reporting is depends on one early decision: SurveyMonkey's managed-package objects, your own fields through custom object mapping, or both. Setting up the connection itself is covered on our SurveyMonkey Salesforce integration services page.
| Reporting need | Managed-package objects | Custom object mapping |
|---|---|---|
| Data shape | A "SurveyMonkey Survey Response" parent per submission, and a "SurveyMonkey Response" child per answered question. | The answers you choose, as fields on a standard or custom object, one record per response. |
| Report type | "SurveyMonkey Responses", once Allow Reports is enabled on both objects. | Your existing report types, so scores appear in reports teams already run. |
| NPS reporting | Child records carry NPS Score, NPS Label and NPS Normalized fields. | A score field you define, with the promoter and detractor logic in a formula you control. |
| Related records | Built-in lookups to Account, Case, Contact, Lead and Opportunity; other objects need extra lookup fields. | Any object you map to, matched by a custom variable carrying the record ID. |
| Best fit | Full submission history and question-level analysis inside Salesforce. | A handful of decision fields (score, reason, follow-up flag) used across many reports. |
Define the metric before you write the survey
Most reporting disputes trace back to a metric nobody defined: which question, which scale, which respondents count, and when. We write each definition down once and design the survey around it.
| Metric | What it measures | Decisions we make with you |
|---|---|---|
| Net Promoter Score (NPS) | Likelihood to recommend on a 0–10 scale. The standard calculation is the share of promoters (9–10) minus the share of detractors (0–6). | Relationship NPS on a fixed cycle or transactional NPS after an event, which contact answers for an account, and whether one account's many answers are averaged or counted individually. |
| Customer Satisfaction (CSAT) | Satisfaction with a specific interaction, such as a closed case or a delivery. | The scale (we keep one scale across every CSAT survey), whether the score is the share of top answers or an average, and which event triggers the survey. |
| Customer Effort (CES) | How easy a task was for the customer, such as getting an answer or completing onboarding. | The exact wording and scale direction (so "higher is better" holds everywhere), and which processes it is measured on. |
| Programme-specific scores | Engagement indices, win/loss reasons, onboarding readiness: anything built from several questions. | Which questions feed the index, how they are weighted, and how partial responses are treated. |
Questions built to be analysed
Closed questions (multiple choice, checkboxes, dropdown, image choice) for anything that will be crosstabbed, because those are the types crosstabs support. One scale per metric, kept identical between surveys so trends and multi-survey analysis compare like with like.
Open text where the reason matters, paired with a scored question, so themes and sentiment can be read against the score.
Segments you do not have to ask for
Custom variables carry known values into the survey link (region, product, account tier, a Salesforce record ID) and appear in exports. Contact custom data appears beside responses and works as a filter in Analyze: 6 fields on every plan, up to 50 on Team and Enterprise plans.
Segmenting by data you already hold keeps surveys shorter and the cut reliable, because nobody has to self-report their region.
Counting the right responses, every time
A metric is only as good as the responses behind it. Duplicates, partial answers and low-quality submissions each move a score, so each gets an explicit rule rather than whatever the default happens to be.
| Issue | What SurveyMonkey does | The rule we set |
|---|---|---|
| Duplicate respondents | Email invitation collectors accept one response per email address per collector. Re-inviting the same contacts from one collector is an Enterprise option. | Which collector types count toward a metric, and how repeat answers from one account are treated. |
| Partial responses | Responses carry a status (completed, partial, over quota, disqualified). The Salesforce integration can sync partial responses page by page if that option is on. | Completed responses only in headline metrics, unless a definition says otherwise; partials analysed separately for drop-off. |
| Low-quality answers | Response Quality flags suspect responses on Advantage, Premier, Team and Enterprise plans. | Flagged responses reviewed and excluded by a saved filter, with the exclusion count reported alongside the metric. |
| Technical fields in exports | Surveys with skip logic or randomization can show Page Order variables, which are not meant for analysis and appear only on incomplete responses. | Excluded from analysis files and BI models by default. |
| Edited surveys | Saved views edited after a crosstab is built do not update the crosstab. | Question or view changes logged against the metric they affect, and dependent reports rebuilt in the same change. |
The right view, for the right person, on a set cadence
A dashboard nobody opens is not reporting. We design reporting around the decisions people make, the rhythm they make them on, and the action expected when a number moves.
| Audience | What they see | Where, and how often |
|---|---|---|
| Leadership | Headline metrics against definitions, trend, response counts and the top themes behind any movement. | Power BI, Tableau or a Salesforce dashboard; monthly or quarterly. |
| Department and team leads | Their own segment, compared with the rest, with significance shown where the plan supports it. | Saved views or a link-shared crosstab in SurveyMonkey; weekly or per cycle. |
| Account and support teams | The latest score and comment on each customer, and open follow-ups. | On the Salesforce record and in list views; as responses arrive. |
| Analysts and researchers | Respondent-level data with custom variables, for deeper modelling. | CSV, XLS or SPSS exports, or a warehouse table; per study. |
Six programmes, six reporting designs
The question each programme has to answer decides where its reporting lives. These are the patterns we design most often.
Post-Support CSAT
Which queues, case types and agents drive low scores? CSAT mapped onto the case in Salesforce, reported beside handle time and reopen rates, with comment themes reviewed weekly by the support lead.
Relationship NPS by Segment
How does loyalty differ by tier, region and product? Account IDs carried as custom variables, NPS trended across cycles in Power BI or Tableau, and the latest score on each account before its renewal review.
Win/Loss Analysis
Why do we win and lose? Surveys tied to closed opportunities, reason codes as closed questions that crosstab cleanly, and results reported against deal size and competitor in Salesforce.
Employee Engagement
Where is engagement moving, and why? Responses tagged with HR attributes such as department and location, the same questions each pulse for multi-survey analysis, and themes read with sensitive surveys kept out of AI features.
Onboarding Check-Ins
Which new customers are at risk early? A short effort and readiness survey at set milestones, scored into one index, and surfaced to the customer success manager on the account.
Market Research
What does the market think of a concept or a price? Studies analysed by specialists in SPSS or Excel from respondent-level exports, kept apart from customer records and operational dashboards.
From raw responses to reporting you can defend
Six phases, each with a deliverable your team keeps. The length depends on how many surveys and audiences are in scope and whether existing surveys must change, and we scope it with you before you sign.
Decisions & Audiences
The decisions survey data should inform, who makes them and how often, plus an inventory of current surveys, reports, exports and BI connections, with your plan, add-ons and data center.
DeliverableReporting requirements and current-state inventory
Metric Definitions
Each metric's question, scale, calculation, inclusion rules, segment keys and owner, agreed with the people who will report on it.
DeliverableSigned-off metric dictionary
Survey Instrumentation
Surveys adjusted for analysis: closed question types, consistent scales, custom variables and contact data for segmentation, Question Bank wording where benchmarks matter.
DeliverableAnalysis-ready surveys and a variable map
Reporting Build
Saved views and crosstabs in Analyze, AI and text analysis settings, Salesforce report types and dashboards, and Power BI or Tableau models, on service accounts rather than personal ones.
DeliverableWorking reports per audience
Reconcile & Validate
The same metric calculated in every place it appears, compared, and any difference explained or fixed before go-live, including how partial and flagged responses are handled.
DeliverableReconciliation log, signed off
Cadence & Handover
Review cadences set up, owners trained to read and explain their metrics, and a runbook for survey changes, plan changes and new reports, with ongoing support on offer.
DeliverableRunbook, training and support plan
What changes when the numbers hold up
The value of survey analytics is not the number of charts. It is meetings that spend their time on what to do, instead of on whose number is right.
One Number per Metric
Every report that shows a score uses the same definition, so departments argue about causes, not arithmetic.
- Written definitions with owners
- Reconciled across every tool
- Exclusions reported, not hidden
Segments on Demand
Results cut by region, product or tier without asking respondents, because the keys travel with the survey.
- Custom variables mapped to CRM data
- Saved views per audience
- Shorter surveys
Comments That Get Read
Open text is themed, tagged and checked, so the reasons behind a score reach the people who can act on them.
- Themes reviewed each cycle
- Tags that match how you work
- AI output labelled and spot-checked
Feedback Beside Revenue Data
Scores sit next to pipeline, cases and renewals in Salesforce, so account teams see sentiment before the conversation.
- Reportable response objects
- Scores on the account and case
- Dashboards teams already open
Less Manual Reporting
Recurring decks and spreadsheets give way to views and dashboards that refresh, so analysts spend time on analysis.
- Connected BI where it fits
- Standard export recipes where it does not
- No laptop-bound connections
Analytics You Can Govern
You can say where survey data and its copies live, which AI features touch it, and who can see each report.
- Data center and copies recorded
- AI features switched on deliberately
- Shared links reviewed
Data and reporting work we have delivered
Neither engagement below is a SurveyMonkey deployment. They are here because the work is the same kind: replacing manual exports with a governed data model, and putting reporting where people already work. Each figure is scoped to the engagement that produced it.
Mid-Sized US Asset Management Firm
A BigQuery data foundation that replaced CSV exports and Excel pipelines for an analyst team.
80% Less manual data preparation per analytics cycle
3 days Cycle start to analysis, down from 2 weeks
Zero Manual CSV uploads after go-live
- Bronze, Silver and Gold layers built with Dataform
- Scheduled pipelines with failure alerting and preserved lineage
US Manufacturer on Salesforce and ServiceMax
Manual meter readings replaced by a native Salesforce utilisation dashboard on each installed product.
- Usage records created automatically when a work order closes
- A Lightning Web Component dashboard with day-wise charts and month, year and custom filters
- Field mapping held in configuration, so tracking changes need no code
- The study reports qualitative results only, so no figures are shown here
More engagements are on our success stories page.
Analytics consultants who know where the data lands
Survey analytics fails less often in the survey tool than in the systems around it: the CRM it should sit beside, the BI model it should feed, and the meeting where it should change a decision.
We start with the decision, not the dashboard
Every report we build traces to a decision someone makes and a metric definition they signed off. Reports with no decision behind them are left out.
We work both sides of the data
Survey design and Salesforce data model in the same engagement, so the keys needed for segmentation exist on both sides. We do that as a Salesforce Partner.
We use what your plan already includes
Analyze, crosstabs and exports answer many questions without a BI licence. We recommend add-ons only when a reporting need requires them, and say so before you buy.
We reconcile before we present
A number shown in two tools is checked in both. Differences are explained in the reconciliation log, not discovered in a leadership meeting.
We bring broader BI experience
Survey data is one source among many. Our data analytics services, Tableau, Power BI and CRM Analytics teams build the models it joins.
SurveyMonkey reporting questions, answered plainly
Metric definitions for the scores your surveys produce, survey changes that make the data segmentable, set-up of analysis in SurveyMonkey (saved views, crosstabs, text and AI analysis), reporting in Salesforce, Power BI or Tableau where it fits, reconciliation of every number across tools, and a reporting cadence with named owners. Ongoing support covers survey changes, new reports and plan changes.
SurveyMonkey's Analyze section handles filtered and compared results, saved views and crosstab reports, which can be shared by link and update as responses arrive. Power BI or Tableau earns its place when you need trends across surveys and periods, survey data next to finance or operations data, or an audience that already works in a BI tool. Both connectors are Enterprise features: Power BI as an add-on for Power Users on Windows Power BI Desktop, Tableau through Extracts.
SurveyMonkey lists crosstabs for Team Premier and above. Analyze with AI, Thematic Analysis, Sentiment Analysis and Response Quality need an Advantage, Premier, Team or Enterprise plan, and statistical significance and multi-survey analysis are listed for Team Advantage and above. Exports are a paid feature. Availability can differ by plan, region and data center, so we confirm features in your account first.
With SurveyMonkey for Salesforce, responses become Salesforce records. Using the managed package, enable Allow Reports on its two response objects and build reports on the SurveyMonkey Responses report type; child records carry NPS score fields. Using custom object mapping, chosen answers land in fields on your own objects and appear in the reports teams already use. Many teams use both: the package for full history, mapped fields for the scores that drive work.
Start with a written definition for each. NPS uses the standard 0 to 10 likelihood-to-recommend question, with the share of detractors (0 to 6) subtracted from the share of promoters (9 to 10). CSAT and CES need decisions on wording, one fixed scale, whether the score is a top-answer share or an average, which event triggers the survey and whether partial responses count. The same definition is then used in every tool that shows the number.
It depends on the survey. According to SurveyMonkey, Analyze with AI sends the survey title, questions and responses to OpenAI, Thematic Analysis uses a self-hosted Anthropic model and OpenAI with common sensitive data types masked, and Sentiment Analysis and Response Quality use SurveyMonkey's own models; the providers do not train on customer data. Primary Admins on some Team and Enterprise accounts can switch features off. We decide per survey, and keep sensitive surveys out.
Yes, on paid plans. Question summaries export as PDF, PowerPoint, XLS or CSV, and individual responses as XLS, CSV, PDF or an SPSS .SAV file, with custom variables included. Exports stay downloadable for 14 days, and surveys above 100,000 responses (or respondents times questions) export as CSV only. For scheduled loads into a warehouse, the SurveyMonkey API is the extraction path.
Next Step
Bring us one disputed number, we will trace it
A first conversation covers your SurveyMonkey plan and data center, the surveys and metrics that matter, and where they are reported today: enough to say what to define, what to rebuild and which tools fit. No obligation to proceed.
Talk to a consultant who works on the survey and the CRM side of the data
