The short answer: dscout has the most complete export of any research platform in this comparison, and it deserves credit for it. You can pull a zipped folder of all photos and videos, a .csv of every response including video transcripts, a .csv of your team's notes, a .csv of comments left on entries, PDFs of all entries, and crosstabs - plus participant demographics, profile pictures, and your message history. The catch is not the format. It is that exporting is a Researcher-level permission, so the person who most needs the data during an offboarding may not be allowed to download it. Sort out who can click Export before you need to.
Credit where it is due
Most articles about leaving a vendor are about what the vendor withholds. This one is mostly not, and pretending otherwise would be dishonest.
Compared with a research repository whose complete export is an encrypted file only the vendor can open, or a video platform whose exports arrive as three disconnected piles, dscout's documented export surface is genuinely broad. The media comes out. The transcripts come out, inside the response .csv rather than as a separate reassembly job. Your researchers' notes come out as their own file. So do comments, which many platforms simply drop.
If you are evaluating dscout, this is a point in its favour, and you should weigh it as one. If you are leaving dscout, it means your migration problem is a different shape than you were braced for.
What actually exports
Per dscout's help documentation on exporting data, a diary study offers these routes:
| What you want | Format |
|---|---|
| All photos and videos | Zipped folder |
| All responses to questions, including transcripts of videos | .csv |
| All notes your team members have made on any entries | .csv |
| Comments left on entries | .csv |
| All entries, sorted alphabetically, oldest to newest, or newest to oldest | |
| Entry responses, researcher notes and highlights, extended participant demographics | Configurable layout toggles |
| Crosstabs | Row and column combinations |
Participant data is a separate export path, covering basic participant demographic data, participant profile pictures, and your message history with participants. Screener exports follow the same pattern, including a PDF of all applications and the same crosstab toggles.
Two things stand out. First, transcripts travel inside the response file rather than as loose artifacts, which removes the single most tedious step in a video-platform migration. Second, researcher notes and highlights are among the data you can toggle into the export layout - so the interpretation layer has a documented route out, which is more than most repositories offer.
The actual constraint: permission
dscout's documentation notes that exporting data from the platform is a Researcher capability, and directs anyone who needs access to it to contact their Account Owner.
In normal operation that is unremarkable access control. During an offboarding it is the whole ballgame, for a reason that has nothing to do with dscout's intentions: the people running a migration are frequently not the people who hold the research seats. A procurement lead, a newly assigned product manager, or an ops contractor brought in to handle the wind-down will typically have no Researcher permission at all. If the original researcher has left the company - which, on a platform you are leaving, is often why you are leaving - the export cannot be run by the person tasked with running it until someone fixes the permissions.
So the first step of a dscout migration is an access audit, not a download.
Why this is urgent in 2026
Two numbers frame the timing. Perspective's 2026 mid-year tooling report surveyed 500 product, research, and CX leaders between February and April 2026 about line-item spend against their 2025 baseline, and found customer research tooling spend down 12 percent year on year, average legacy CXM seat counts down 38 percent, and 22 percent of enterprise buyers beginning a full migration off their incumbent platform. Treat it as one vendor's read on its own market rather than a census, but consolidation is clearly under way, and seat reductions are exactly the event that strands an export permission with someone who no longer works there.
Second, the method itself is being re-evaluated. Maze's Future of User Research Report 2026 found 69 percent of respondents using AI in at least some research projects, up 19 percent year on year. Diary studies are among the most expensive instruments in research - weeks of elapsed time, per-participant incentives, and heavy analysis - so teams are increasingly asking whether a longitudinal study is the right tool or just the familiar one. That question determines how much history is worth migrating.
What does not travel, even with a clean export
This is the part specific to diary research, and it is structural rather than a vendor limitation.
A diary study's value is not in any single entry. It is in the sequence - what a participant said in week one versus week four, how a frustration escalated or resolved, which behaviour persisted after the novelty wore off. That longitudinal structure lives in the relationship between entries and in the study design that produced them: the cadence, the prompts, the reminders, the eligibility rules.
A .csv of responses preserves the rows. It does not preserve the instrument. Nothing you can download tells the next platform how the study was scheduled, or re-creates the per-participant timeline as a first-class object rather than as a sortable date column.
Practically, that means two things:
- Export the study design, not just the data. Document the cadence, the entry prompts, the screener criteria, and the incentive structure as a written artifact. It is not in any export file, and it is the thing you would otherwise have to reinvent.
- Accept that old diary data ages badly. A diary study of a product that has shipped four releases since is reference material, not evidence. Migrate the write-ups and the design; archive the raw entries.
For method background, the diary study guide covers how cadence and prompt design drive data quality, which is exactly the documentation you need to recreate.
The extraction sequence
Step 1 - Audit export permissions first. Establish who currently holds Researcher access, confirm those people are still at the company and still have seats, and get the Account Owner to grant whatever is missing. Do this in week one. Everything else is blocked on it.
Step 2 - Export study data per study. Take the zipped media, the response .csv with transcripts, the notes .csv, and the comments .csv. Four files, one pass, nothing to rejoin by hand.
Step 3 - Toggle highlights and extended demographics into the layout. The defaults are not the maximum. Turn on researcher notes and highlights and extended participant demographics before you export, not after.
Step 4 - Take the PDFs for studies that get reread. Sorted oldest to newest, a PDF of all entries is the closest thing to a readable longitudinal record, and it is the format a future colleague will actually open.
Step 5 - Export participant data and message history separately. Demographics, profile pictures, and your message history with participants are a different path. The message history matters more than teams expect, because it documents how you handled dropout and non-compliance.
Step 6 - Write down the instrument. Cadence, prompts, screeners, incentives. The part no export contains.
Step 7 - Run the replacement in parallel before cancelling. For notice periods and contract timing, see how to switch user research platforms. On cost, dscout pricing covers what the contract actually includes, and dscout alternatives covers the field. If a data protection review is part of the exit, the research vendor DPA review guide covers the clauses governing return and deletion.
What to require from the replacement
- Is export a right or a permission? The dscout lesson generalises: confirm who can export without filing a request, and what happens when that person leaves.
- Do transcripts ship with responses? Transcripts inside the response file rather than beside it saves the single most tedious migration step. Hold the next vendor to the standard dscout already meets.
- Do notes and highlights have a structured route out? Not just a PDF.
- Is there an API? UI-only export caps migration speed at human clicking speed.
- Can you reproduce the instrument from the platform? If the study design is not exportable or documentable, you will rebuild it from memory next time.
Why teams move to Koji
The honest case for Koji here is not that dscout's export is bad - it is that a diary study is a slow, expensive way to answer most of the questions teams point it at, and Koji answers them in a different timeframe.
Koji runs AI-moderated voice interviews that participants complete on their own schedule, with no moderator to roster and no multi-week field period to coordinate. Thematic analysis runs automatically across every transcript, and the output is a shareable report in one click. For the large class of questions where a diary study was chosen mainly because you needed depth from a lot of people, that is the difference between weeks and hours - and no moderator in the loop also means no moderator bias shaping the answers.
On portability, Koji is designed so the export is not an event. Study data exports to CSV as an immediate download with participant metadata, session data, AI summaries, and quality scores in one file, and the API returns the full structured analysis as JSON - the same data Koji uses internally to build reports. Transcripts are accessible per participant. There is no encrypted archive format and no emailed link that expires. The exporting research data guide documents every route including the API endpoint, and API authentication covers the credentials.
The instrument problem is handled differently too. A Koji study is built from six structured question types - open_ended, scale, single_choice, multiple_choice, ranking, and yes_no - each with a stable question ID, so the study design is itself structured data rather than a configuration you have to describe in a handover document. Open-ended answers get AI follow-up probing; scale questions become distributions; rankings become average positions. Koji also ships methodology frameworks including Mom Test and jobs-to-be-done, so a discovery study starts from a question set that probes mechanisms rather than opinions. The structured questions guide covers designing for all six types.
On trusting the analysis: the Condens State of AI in User Research Analysis survey of 332 practitioners found 71 percent agreeing AI lets them analyse data significantly faster, while 71 percent also agreed validating AI output still takes significant time and only 21 percent trust it with minimal review. The Lyssna Research Synthesis Report 2025, across 300 professionals, found 60.3 percent naming time-consuming manual work as their biggest synthesis challenge. Koji's answer is traceability rather than trust: every theme links to the quotes and participants behind it, and each interview carries a one-to-five quality score so thin sessions can be filtered before they reach the report. See Koji vs dscout for the direct comparison, and the thematic analysis guide for how the coding works.
Before you renew a diary study contract, test the faster path. Run a Koji study this week, export it on day one, and compare what you get against the weeks a diary study would have taken.
Frequently asked questions
Can I export all my dscout data?
Largely, yes - dscout's export surface is one of the more complete in the category. A diary study exports a zipped folder of all photos and videos, a .csv of all responses including transcripts of videos, a .csv of team members' notes on entries, a .csv of comments left on entries, PDFs of all entries, and crosstabs. Participant demographics, profile pictures, and message history export on a separate path.
Who is allowed to export data from dscout?
dscout's documentation describes exporting as a Researcher capability and directs anyone who needs access to contact their Account Owner. This matters during an offboarding, because the person handling a migration is often in procurement or ops rather than research. Audit who holds Researcher access, and whether they still work there, before you plan any download.
Do dscout exports include video transcripts?
Yes. Transcripts of videos are included in the .csv of responses to questions rather than arriving as separate files. That is a meaningful advantage over platforms where video, metrics, and transcripts download separately and have to be rejoined by hand, and it removes the slowest step from the migration.
Do researcher notes and highlights come out?
They have a documented route out. Notes your team made on entries export as their own .csv, and researcher notes and highlights can be toggled into the export layout along with extended participant demographics. Check those toggles before exporting, because the default layout is not the maximum.
What does not survive a dscout migration?
The instrument and the sequence. No export file contains the study design - the entry cadence, the prompts, the screener criteria, the incentive structure - and a flat response file preserves rows rather than per-participant timelines as first-class objects. Write the design down as a separate document, and treat old raw entries as reference material rather than live evidence.
Is it worth migrating historical diary study data?
Partially. Diary data about a product that has shipped several releases since is context, not current evidence. The durable assets are the synthesis write-ups, the study design, and the participant message history that records how you handled dropout. Archive the raw exports for completeness, but do not spend migration effort reconstructing them in a new platform.
