đ Segmentation analysis is available on every OpinionX plan, including Free.
Segmentation analysis on OpinionX
Segmentation analysis means splitting your survey results by who answered, so you can see how different groups of participants think. It's sometimes called subgroup analysis, cohort analysis or crosstab analysis. Instead of one overall answer, you see what your Enterprise customers ranked first, how designers' priorities differ from engineers', or which group would pay the most.
OpinionX has four segmentation tools, and each one answers a different question. This guide explains what segmentation is, which tool to use, where segment data comes from, and how to plan your survey so the segments you need are there when your results come in.
-- -- --
What is segmentation analysis?
A segment is a group of participants who share something, like a job role, a pricing plan or a country. Segmentation analysis recalculates your results for each segment, so you can see where groups agree and where they differ.
There are two broad ways to define segments:
By who people are â Segments based on facts about each participant, like role, company size, plan or country. These are sometimes called demographic or firmographic segments, and they usually come from a survey question or from data you already have.
By what people want â Segments based on participants' priorities and preferences, found by grouping people who answered the same way. This is called needs-based segmentation, and it often reveals groups that cut across roles and plans. Our guide to needs-based segmentation explains the approach in depth.
-- -- --
Why segment your results?
An overall result is an average, and an average can hide groups that want opposite things. Imagine a feature prioritisation survey with a MaxDiff Analysis block of 12 features, sent to 400 customers. Overall, "Bulk export" ranks 4th. Filter to customers on your Enterprise plan and it ranks 1st; filter to Free users and it drops to 10th. The overall ranking described neither group.
Every answer in an OpinionX survey belongs to one participant record, so the same segments work across every block in that survey. You can see what Enterprise customers ranked in your MaxDiff Analysis and what they'd pay in your Van Westendorp question, from the same people, without matching data between studies. OpinionX runs the analysis for you automatically, no manual work required.
-- -- --
Which segmentation tool do I need?
Start from the question you're trying to answer:
Your question | Tool | What it does |
What does this one group think? | Filter by Segment | Recalculates every result on the page for one segment |
How do these two groups differ? | Compare Tab | Shows two segments side by side, with the differences between them |
How does every group score on everything? | Segments Tab | Crosstabs each result by many segments at once, as a heatmap |
Are there groups I didn't know to look for? | Cluster Tab | Groups participants who answered similarly, using AI-powered clustering |
Filter by Segment
Build a segment from one or more conditions, like "Role is Product Manager" and "Plan is Enterprise", and every result on your Results Overview recalculates for just those participants. It's the quickest way to answer "what does this group think?" Read the Filter by Segment guide.
Compare Tab
Define two segments and see their results side by side, block by block. For ranking blocks, it shows how far each option's rank and score moved between the two groups, so the biggest differences stand out. Read the Compare Tab guide.
Segments Tab
The Segments Tab is where OpinionX shows crosstabs: tables with one row or column per segment and a heatmap that makes high and low scores easy to spot. Pick up to 30 segments and see how each one scored every option, so you can scan for the group that disagrees with everyone else. Read the Segments Tab guide.
Cluster Tab
The Cluster Tab finds needs-based segments for you. AI-powered clustering uses K-means, a machine learning algorithm, to group participants whose answers to your ranking blocks were similar. Each cluster card then shows what that group cared about most and who's in it. Read the Cluster Tab guide.
-- -- --
Where segment data comes from
You can build segments from four kinds of data.
Answers to your survey questions
Participants' answers to these blocks can define segments:
Multiple Choice, the most common segmentation question (eg. "What's your role?")
Rating Scale (eg. everyone who rated their satisfaction 1 to 3)
Number Response (eg. companies with more than 50 employees)
Matrix Grid (eg. everyone who agreed with one statement)
Gabor Granger interest (participants who were interested in the product or not)
Data you already have about participants
You don't have to ask participants for things you already know. Add your own customer data, like plan, industry or account age, on the Enrichment Tab, then match it to participants with Hidden Fields, Unique Links or Email Invites. Fields set to the Categorical data type become segments.
Data OpinionX records for you
Language â which language each participant took a multi-language survey in.
Finish Page â which Finish Page they reached.
Participant Status â whether they completed, left partway, were screened out or were over quota (see Participant Status).
Date Completed â when they finished, so you can compare results before and after a launch.
Participants â hand-picked individual participants.
Clusters
Once you've run clustering on the Cluster Tab, each cluster becomes a segment too, so you can filter, compare and crosstab your needs-based segments just like any other group.
Answers to identifier blocks, like Email Address or Username/ID, and to Text Response questions can't be used as segments, because almost every answer is different. To group people by something they type, ask it as a Multiple Choice question instead.
-- -- --
Plan your segments before you launch
Segments can only use data your survey collected, so decide which groups you'll compare before you share the link:
List the comparisons you'll need to make your decision, like "Enterprise vs everyone else" or "each job role".
Check you'll have the data for each one. If you already know it, pass it in through the Enrichment Tab or Hidden Fields. If you don't, add a Multiple Choice question.
Keep categories consistent. Imported category values are case-sensitive, so "Pro" and "PRO" become two separate segments. Tidy your data before you import it.
Make sure each segment fills up. Use Quota Management to set a target number of participants for each group you plan to compare.
How many participants does each segment need?
Each segment is a smaller sample than your whole survey, so its scores move more with every answer. Before you read much into a difference, check how many participants are in each segment. The Cluster Tab flags any cluster with fewer than 15 participants as a Small Sample for this reason.
If a comparison matters to your decision, size each segment the way you'd size a whole study. The sample size calculators in the MaxDiff Analysis and Conjoint Analysis guides help you work out that number.
-- -- --
Segments across OpinionX
Saved Segments â Save a segment once and load it again in Filter by Segment, the Compare Tab and the Segments Tab. See Saved Segments.
Participants Tab â The Filter your data button on the Participants Tab uses the same segment builder, so you can see the individual people behind a result.
Exports â If a segment filter is active when you export your results, the export uses that segment.
Segmentation vs branching
Segmentation shows everyone the same questions and splits the results afterwards. Branching changes which questions each person sees while they take the survey. Many surveys use both: branching to show each group the questions that apply to them, and segmentation to compare how the groups answered.
-- -- --
Frequently asked questions
Is segmentation analysis free on OpinionX?
Yes. Filter by Segment, the Compare Tab, the Segments Tab and the Cluster Tab are on every plan, including Free.
Can I segment by data I already have about my customers?
Yes. Add it on the Enrichment Tab as a Categorical field, then match it to participants with Hidden Fields, Unique Links or Email Invites.
Can I segment by an open-text answer?
Not currently. Text Response answers can't define segments. If you'll want to group people by something, ask it as a Multiple Choice question.
Does segmentation work with multi-select Multiple Choice questions?
Yes. Segments work with both "Pick One" and "Pick Many" questions.
How many segments can I compare at once?
The Compare Tab shows two segments side by side. To see many segments at once, use the Segments Tab, which handles up to 30.
Does applying a segment change my data?
No. A segment only changes which participants' answers are shown. Clear it and you're back to your full results.
-- -- --
Still stuck?
Ask us in the live chat in the bottom-right corner of this page, or âĄď¸ book a survey setup call. The Analyze plan includes 90 minutes per year of prepaid consulting with an OpinionX expert, and Accelerate includes more. See Expert Assistance.
