Ipsos is one of the three largest custom market research firms in the world, and unlike most of its competitors it is publicly listed, which means it tells you exactly how its business is going. Koji is an AI-native research platform that runs moderated interviews with your own customers and returns coded themes the same day. This comparison uses Ipsos published numbers rather than estimates, because they are more interesting than any vendor claim.
Short answer: choose Ipsos when you need a named human accountable for a complex multi-country study, regulated or public-sector work, or methodological defensibility that will be challenged by someone outside your company. Choose Koji when you need to understand your own customers quickly and repeatedly, and the cost of a six-week loop is a decision made without evidence. The two are complements far more often than substitutes.
What Ipsos reported for 2025
Ipsos published its full-year 2025 results in February 2026. The headline figures:
- Revenue of 2,525 million euro, up 3.4 percent in total.
- Organic growth of 0.6 percent.
- A scope effect of 5.8 percent, driven mainly by the acquisitions of The BVA Family and infas, and a negative 3.0 percent currency impact.
- Operating margin of 309 million euro, or 12.3 percent of revenue, and 12.8 percent at constant scope.
- Regionally, EMEA revenue reached 1.245 billion euro and grew 12 percent in total, with the Americas down 3.4 percent and Asia-Pacific down 4.3 percent.
- Ipsos launched a strategic plan named Horizons in January 2026 and guided to 2 to 3 percent organic growth for 2026 with a margin similar to 2025.
Ipsos described the year candidly. Growth was affected by an uncertain political context across several geographies and tighter government budgets, which hit Public Affairs particularly hard. Industry press reported the result as revenue growth that was, in the company's own framing, below its ambitions.
The number that should shape your buying decision
Look at the gap between 3.4 percent total growth and 0.6 percent organic growth.
Almost all of the difference is the 5.8 percent scope effect, which is another way of saying the growth was purchased. One of the best-run firms in the industry, with genuine scale and a strong brand, grew its existing business by six tenths of one percent.
That is not a criticism of Ipsos management. It is a structural signal about the delivery model itself, and it applies to every full-service firm, listed or not.
Here is the mechanism. In a services business, revenue is coupled to hours. A custom research project consumes researcher time in fixed proportion to its size: scoping, questionnaire design, moderation, coding, charting, debriefing. To grow revenue you must add people, and adding people adds cost at nearly the same rate, which is why the operating margin sits at 12.3 percent rather than the 70 or 80 percent gross margins software businesses report. The model is sound, but it does not compound. Each new project starts from zero.
Software decouples those two things. A platform that runs 40 interviews costs almost exactly what it costs to run 4, which is why the marginal cost of your next question approaches nothing. That is the entire economic argument for AI-native research, and Ipsos own income statement demonstrates it more convincingly than any vendor benchmark.
The practical consequence for you as a buyer: hour-coupled research must be rationed, and rationed research gets asked late. Teams save their agency budget for the big decisions, which means the medium-sized questions never get researched at all. They get answered in a meeting by whoever is most confident.
What Koji actually does
Koji writes an interview guide from your brief, then an AI moderator conducts the interviews itself by voice or by text, in the participant's own language, probing where a respondent says something the guide did not anticipate.
- AI-moderated voice and text interviews running in parallel, at any hour, with no scheduling.
- Six structured question types in the same study:
open_ended,scale,single_choice,multiple_choice,rankingandyes_no. A single Koji study returns an NPS distribution and the reasons behind each score. The structured questions guide covers how these aggregate. - Automatic thematic analysis, with descriptive and in-vivo codes grounded in verbatim participant words, then clustered into a canonical codebook across the whole study. This is the workflow described in our thematic analysis guide, performed automatically instead of over a fortnight.
- One-click reports where every theme traces back to the transcript line that produced it.
- No moderator bias and no moderator fatigue. Interview 40 is conducted exactly like interview 1.
Ipsos vs Koji: which one for which job
| The job to be done | Better fit | Why |
|---|---|---|
| Nationally representative quant with a defended sampling frame | Ipsos | Panel infrastructure, published methodology, external defensibility |
| Public affairs, polling and government work | Ipsos | Specialist practice, regulatory familiarity, public credibility |
| Multi-country study needing local field teams | Ipsos | On-the-ground presence and local methodological expertise |
| Understanding why churn spiked last quarter | Koji | Your actual churned accounts, results in days, themes coded automatically |
| Message and positioning testing before a launch | Koji | Cheap enough to iterate several rounds instead of buying one verdict |
| Continuous discovery for a product team | Koji | Per-conversation pricing makes weekly research economically sane |
| Win-loss interviews with recent buyers | Koji | Your CRM list, not a panel, and no scheduling friction |
| Benchmarking against category norms built over decades | Ipsos | Historical comparability cannot be generated by any platform |
Where Ipsos is clearly the right buy
- Public affairs and polling. This is a specialist discipline with its own standards, and Ipsos is one of a handful of firms that does it credibly at national scale.
- Studies that will be publicly challenged. When the output faces a regulator, a court, a journalist or a hostile board, you want a named firm that will defend the methodology in its own voice.
- Genuine multi-country fieldwork with local nuance. Local teams who understand why a question lands differently in Jakarta than in Lyon are a real asset.
- Complex methodology you do not want to own. Conjoint, MaxDiff, segmentation with a modelling layer, and similar work benefits from specialists.
- Risk transfer. An agency accepts responsibility for the study being right. That is a product in itself, and some organisations are correct to buy it.
For the wider landscape, see our comparison of the best market research companies in 2026 and the related looks at Koji vs Kantar and Koji vs YouGov.
Where Koji wins
The second latency
There are three separate clocks in any study: time to field, time to understand, and time to a decision. Automated survey products, Ipsos.Digital included, have already compressed the first one. Almost nobody has compressed the second. A study that fields in 48 hours and returns an analysed debrief a month later has moved the fast clock and left the slow one untouched. Koji codes interviews as they arrive, so the report is substantially written when fielding closes.
Cost per question, not cost per project
Koji charges credits: a text interview costs 1 credit, a voice interview 3, and a report refresh 5. At 1 euro per credit, 40 voice interviews is 120 credits. The point is not that it is cheaper than an agency, which is obvious. The point is that it is cheap enough to be iterative: run a study, read it, discover you asked the wrong question, and run the better one tomorrow. That loop is how research actually generates understanding, and hour-coupled pricing makes it impossible.
Your customers, not a population
Ipsos samples populations expertly. But the accounts that churned last quarter are not in any panel, and neither are your trial users who never activated. Those studies need your list. This is where the jobs-to-be-done interview method and structured customer validation work pay off fastest, because you are talking to people with a real invoice and a real renewal date.
Depth that a questionnaire cannot reach
A survey commits every question before the first answer arrives. An interview commits the guide and improvises the rest. If you are deciding between the two instruments for a given question, our surveys vs interviews guide is the practical decision tree, and the voice vs text mode comparison covers which modality suits which audience.
Evidence that accumulates
Agency deliverables arrive as decks and tend to die in a drive folder. Koji studies stay queryable, which matters more than it sounds: the value of research compounds only if you can find last quarter's finding when this quarter's question arrives. Our research repository guide covers the discipline either way.
What each one costs
Ipsos does not publish a rate card, and pricing depends entirely on markets, sample, method and service level. Custom full-service work is quoted per project.
Koji publishes everything:
- Insights: 29 euro per month, including 29 credits.
- Interviews: 79 euro per month, including 79 credits.
- Annual: 290 euro and 790 euro, which is two months free.
- Extra credits at 1 euro, with packs at 50 for 50 euro, 100 for 95 euro, and 250 for 225 euro.
- Enterprise: custom.
- 10 free credits on signup, no card required.
And a term no services contract offers: a quality gate means only conversations scoring 3 or above consume a credit. A junk interview is free.
How to decide
Ask who has to believe the answer.
If the audience is external, adversarial, regulatory or public, buy the firm whose name carries the defence. That is Ipsos territory and no platform changes it.
If the audience is your own team, and the question is why your customers do what they do, then speed and iteration beat brand. Do the work yourself, in days, with the people who actually pay you.
Most serious insights functions in 2026 end up with both: a firm for the studies that must be defended, and a platform for the ninety percent of questions that simply need answering before the decision is made.
Related reading: for the syndicated data side of the full-service model, see NielsenIQ vs Circana vs Numerator and syndicated data vs custom research.
Frequently Asked Questions
Is Koji an alternative to Ipsos?
For part of the portfolio, yes. Koji replaces the custom qualitative work that makes up a large share of agency spend: churn diagnosis, discovery, message testing, win-loss, pricing objections and concept feedback. It does not replace public affairs polling, nationally representative quant with a defended sampling frame, or complex modelled methodologies. Most teams that adopt Koji keep an agency relationship for the studies that must survive external challenge.
What did Ipsos actually report for 2025?
Revenue of 2,525 million euro, up 3.4 percent in total but only 0.6 percent organic, with a 5.8 percent scope effect from acquisitions including The BVA Family and infas and a negative 3.0 percent currency impact. Operating margin was 309 million euro, or 12.3 percent. EMEA grew 12 percent to 1.245 billion euro while the Americas declined 3.4 percent and Asia-Pacific declined 4.3 percent. Ipsos launched its Horizons plan in January 2026 and guided to 2 to 3 percent organic growth for the year.
Why does organic growth matter when choosing a research supplier?
Because it shows whether the underlying delivery model still compounds. Services revenue is coupled to hours, so growth requires adding people, which is why margins sit near 12 percent rather than software levels. For a buyer this translates into a practical constraint: hour-coupled research has to be rationed, and rationed research gets commissioned late or not at all. A platform where the marginal cost of the next study is near zero changes which questions are worth asking.
Can Koji handle multi-country research?
Yes for interview-based work. The AI moderator conducts interviews in the participant's own language, and studies run in parallel across time zones without local scheduling. What Koji does not provide is local field teams, in-person work, or country-specific panel infrastructure. If your study needs nationally representative samples in eight markets with locally defended methodology, that remains agency territory.
How does Koji handle analysis compared to an agency team?
Every open-ended answer is coded during the study with descriptive and in-vivo labels grounded in the participant's own words, then near-duplicate codes are clustered into one canonical codebook across all interviews. Structured questions aggregate into charts in parallel. The output is a report where each theme links back to the transcript that produced it, so a skeptical stakeholder can check any claim in seconds rather than trusting a summary slide.
Is an AI moderator credible for senior B2B respondents?
In practice, yes, and often more so than expected. Senior respondents value not having to find a calendar slot, and an AI moderator that runs at 9pm on their schedule removes the largest single barrier to B2B participation. The consistency also helps: every participant gets the same core questions, so the comparison across accounts is clean. Where it still struggles is highly adversarial or politically delicate conversations, which remain better served by an experienced human.
Answer the questions you have been rationing
The research you skip is rarely the big study. It is the medium question that did not justify a statement of work, and got answered by opinion instead.
Start free with 10 credits. Write a brief, launch a study with your own customers, and read coded themes today. No scoping call, no procurement cycle, no card.
