Product · deepsight cloud Understand numbers and open-ended responses together.
Upload existing data or collect new responses, analyze open-ended answers together with your ratings, and share the results as a report or in PowerPoint.
Watch the tour · 35 seconds
Illustrative example · Read the content
An illustrative brand study runs through deepsight cloud: start a survey or import data, ask an AI follow-up, analyze topics and sentiment, and create reports. Export to Excel or PowerPoint, or share online. The aggregate NPS of +31 comes from 49% promoters minus 18% detractors; individual NPS responses use a 0–10 scale.
Why a seven – and not a ten? The reasons are in the open-ended responses.
Thousands of them. But before they turn into results, the data moves from tool to tool – export after export, week after week.
The challenge: A questionnaire here, spreadsheets there
Responses are collected in the survey tool and analyzed in Excel, while the comments sit in a separate list – apart from the ratings.
With deepsight: Numbers and reasons together
Open-ended responses stay linked to ratings and attributes. Group comparison shows which topics dissatisfied respondents mention, for example.
Go to group comparisonThe challenge: Coding by hand
Reading response after response and assigning codes – with thousands of responses, that takes weeks. Often there is only time for a sample.
With deepsight: Assigned in minutes, not weeks
Topic analysis assigns the open-ended responses – to new categories or to your codebook. You review the assignments instead of coding everything yourself.
Go to topic analysisThe challenge: Charts rebuilt in PowerPoint
For every presentation, charts are redrawn and quotes are copied out of the raw data file.
With deepsight: From finding to report
Pin charts and quotes right where you analyze and put them together in a report – to share via link or to keep editing in PowerPoint.
Go to reportingThe challenge: Starting over, week after week
With tracking studies, the analysis starts from scratch with every wave. Whether a topic is growing or shrinking only becomes clear after matching it up by hand.
With deepsight: Waves as a time series
In the time series, deepsight links topics with the same name across waves and suggests further matches for you to confirm.
Go to data and follow-up waves
Topics, shares and the responses behind them
deepsight organizes open-ended responses by topic and connects them with the rest of your data. Select a topic to see the responses assigned to it.
“Waiting three weeks for a sofa is too long.”
“The delivery arrived a day earlier than announced, great.”
Illustrative data.
Three ways to get data into the cloud
Upload existing data
An Excel or CSV file with open-ended responses, ratings and attributes. You can keep using your current survey tool.
Run a new survey
Draft a questionnaire from your brief, review it and collect responses in the cloud. We activate surveys for your organization as part of the introduction.
Go to surveysConnect a source
Fetch reviews from the Apple App Store and Google Play via connectors and analyze them as monthly waves. To do so, you add your App Store Connect or Google Play Console access. Special system integrations are available as custom projects.
Example: reviews & feedbackRecurring analysis: Add follow-up waves and newly fetched reviews to a time series, with a separate one for each source. Align topics across waves and compare how they develop.
Wave comparison and review connectors are included in every paid plan – see the plan comparison.
What the cloud brings to your analysis
Prepare data
Review and adjust results
Use results
Illustrative example · Read the content
A fictional customer survey for a mobility app: the group comparison shows more negative responses about vehicle availability from frequent riders (71%) than from occasional riders (48%), with a significance flag. Asked “What do dissatisfied frequent riders criticize?”, the assistant answers with a chart (41% vs. 18% of detractors mention availability) and two original responses. The answer is pinned for the report.
Compare groups, ask questions, show the evidence
Compare groups and ask questions about your analysis – with the original responses as evidence.
- Group comparison: cross-tabulate topics and ratings by attributes from your data, with significance indicators and optional weighting.
- Assistant: ask questions in your own words; depending on the question, answers include charts and original responses as evidence.
- For the report: pin findings and use them in your report later.
Report, PowerPoint or raw data
Illustrative example · Read the content
In a fictional customer barometer for a mobility app, pinned results – a chart, a topic overview and an original response – fill a report. Exporting as a presentation creates editable PowerPoint slides, and the design can be switched. Sharing creates a link that lets recipients open the report without an account.
According to the plan comparison, reports, customizable slides and PowerPoint export are included in every paid plan. Share reports via a link – recipients do not need an account. Data can be exported to Excel or CSV.
You only need slide factory if you want to turn tables from other programs into slides based on your template.
A look inside deepsight cloud
Understand what is behind a topic.
Read the original responses alongside sentiment and topic metrics.
What your data can tell you
- Which topics are praised, and which are criticized?
- Do regions, plans or customer groups differ in what they raise?
- What has changed since the last wave, and which topics are new?
- Which comments contain a complaint, a question or an explicitly stated intention to cancel?
- What do respondents write when they answer in another language?
- Which quotes back up a finding for your presentation?
Collect. Analyze. Explore. Report. All in one project.
Whether you run the survey in the cloud or upload data from your current tool.
From raw data to a reviewed analysis
- Prepare the data
Detect empty, unintelligible or content-free responses and filter them out before the analysis. They only stop counting toward your text allowance once you have completed the check.
Data quality check - Topics, sentiment, intent
Topic analysis assigns the responses to topics. If you wish, you can add the sentiment for each topic and the intent of a comment, such as praise, a question or a complaint.
Sentiment analysis - Review and refine
Read samples for each category, correct assignments and have the residual “Other” category searched again. Every change is saved as a topic version.
Topic refinement - Build an overview
Bring topics, ratings and groups together in a dashboard, move from there to the comments and share it via link.
Dashboards
Good to know: If you choose, anonymization masks names, email addresses or phone numbers in responses before the analyses run. If translation is switched on, texts are translated before masking. Analyses started directly from a survey run without anonymization. Anonymization doesn’t catch every detail – check the results before sharing them.
What you end up with
The AI suggests. You decide.
You can check any assignment against the original response, and you choose which assistant answers go into your report.
What matters in the analysis
Base and multiple mentions
Shares and metrics such as the NPS refer to the analyzed responses in each group; anything the data quality check filters out is not part of the base. Because one response can be assigned to several topics, the shares often add up to more than 100%.
Put differences in context
If you wish, significance indicators show whether a group differs noticeably from the rest of the sample in statistical terms; they require at least 30 responses per group. Values below a fixed base of 10 responses are flagged or hidden. The comparison doesn’t show why a group answers differently.
Scales and boxes
For scales, deepsight shows the mean as well as top-2 and bottom-2 shares; you decide which end of the scale counts as good.
The same settings for every wave
Save the settings for your study as an analysis profile. It is versioned and can be reused with your codebook for every follow-up wave.
Questions about deepsight cloud
A general-purpose chat answers questions about the texts you give it. deepsight cloud assigns the responses in a dataset to topics, links them with ratings and attributes, and saves the results in your project. Shares and significance indicators are calculated, not written by a language model.
An analyzed text is one open-ended response, comment or review; closed-ended responses and quantitative tables don’t count. A new completed analysis run can count the same texts again. If you have completed the data quality check, filtered-out rows don’t count; without it, the number of rows in the dataset counts. The pricing page shows the allowance for each plan.
deepsight cloud runs on Microsoft Azure in the EU, and for AI features we use Azure OpenAI. Microsoft does not use your data to train AI models. DeepL or Google are only used for translation if you choose these services. We sign a data processing agreement under Art. 28 GDPR, and a list of the service providers we use is available on request. Encryption, retention periods and support access are described on the Security page.
The deepsight cloud interface is currently in German. You can still analyze responses in other languages: a built-in translation (AI-based, or DeepL or Google if you choose) supports the analysis.
Yes. Get to know the solution in a personal demo. Afterwards, you receive test access. We agree the scope of the test with you, based on your project.
Related to deepsight cloud
See the cloud with your own data
We will show you the explorer, the review and the export using an example from your field.