October was busy: six releases shipped, and three of them deserve the spotlight:
- Chat exports to PowerPoint and Word: Chat now hands you finished files at the end of a study, so you can skip the copy-paste marathon.
- Shared survey inbox: Your whole team can work on the same survey without stepping on each other’s toes. Its sidekick, questionnaire review, catches problems before a single line gets programmed.
- Usage-based pricing: This is the one your finance team has been waiting for. You pay for what you actually use, and the account pages got a makeover to match.
Here’s the short version, with links if you want to go deeper.
Decks and documents, out of Chat
Ask Chat to build a PowerPoint or a Word document from a survey’s reports, uploaded documents and your account’s template. It proposes a plan and the sources it intends to use, you approve it or hold it, and it returns a version you can preview page by page, download, and reopen later.
A generated deck in the document preview, before download.
Charts and tables come out native and editable rather than pasted as pictures, so a client can change a label without coming back to you. Account templates keep their design, existing tracking decks keep their layout, significance labels you supply are carried through, and appendices stay readable within the approved slide count.
Generation survives a reconnect, can be cancelled part-way, and keeps its version history. One retirement to note: direct insight-to-PowerPoint export has gone, and editable documents are created through Chat instead.
Learn more: Chat · Editable PowerPoint charts in MX8 Labs
More than one person on a survey
A survey now has an Inbox. Start a thread about a change, mark it blocking if it has to be settled before fielding, assign an owner, and archive it once it is resolved. There are views for unread items and for blockers, and unread is personal: a colleague reading a thread does not mark it read for you. Thread titles are generated from the messages.
Feedback from testing arrives in the same place. A reviewer selects Comment while a question is on screen, and comments about that question group into one thread that keeps its link to the question through later edits.
Survey activity records saved changes, who made them and when. Open a change to see which settings it touched and undo it where that is possible; later edits that remain compatible are preserved, and where they are not, you see the conflict before deciding.
A saved change in Survey activity, with its undo.
Two people can also save the same survey at once. Independent edits are merged, and overlapping ones come back as a conflict you can act on rather than a silent overwrite.
Learn more: Collaborating with others on a survey · Testing and reviewing your survey
Catch questionnaire problems before programming
Paste in a client’s questionnaire and it is reviewed before anything gets built. Conflicting instructions, missing rules and requirements the platform does not support are flagged against the passage that caused them, each with a recommended change you can apply or dismiss. Survey creation has a project step as well, so a new study lands in the right place.
Questionnaire review, with each finding linked to the passage it came from.
Question groups collect related questions under one heading, or repeat the same set of questions across named rows — one row per person in the household, per brand, per occasion. Each question inside the group keeps its own options, validation and reporting identity.
A repeated question group, one row per person.
Choice and grid groups are the smaller sibling: non-selectable headings inside a single question’s options, grid rows or numeric-grid columns, so a long list is easier to read. The headings never become response values or reporting rows.
Also in the questionnaire: explicit Don’t know answers across eligible text, numeric and grid questions; Other, please specify rows in select, multi-select, rating and numeric grids, keeping the row label alongside the answer; optional grid lines on wide desktop grids; and editable starting values for numeric grids. Guided walkthroughs cover creating projects and surveys, exploring a survey and setting up respondent sources, and can be replayed from the user menu or from Chat.
Learn more: Creating a new survey · Question groups · Choice and grid groups
Video that asks about the moment
Dial-test reactions are now usable downstream. Rises, falls, stable periods, peaks and troughs can each be referenced in follow-up questions and in reporting, so you can ask someone why the line dropped and show them the frame it dropped on — protected video thumbnails can be retrieved at a chosen timestamp.
A follow-up question built from a dial reaction, with the frame it refers to.
Video can also be played as an authored segment, with optional start and end times, while full-video playback and the controls protecting secured video stay as they were.
Learn more: Dial testing with MX8 Labs · Secure video player · Video dial testing with personalized follow-up questions
Also new this month
- A redesigned survey experience across chat and traditional forms, with refreshed question layouts, controls, media presentation and loading transitions on desktop and mobile.
- Themes and respondent-facing interface copy can be edited with a live respondent preview, and translated independently of the questionnaire. Chat and traditional layouts can carry a client logo.
- Opportunity-to-see lift reports, for campaigns where exposure records are incomplete: build a score from the media plan and respondents’ media habits, then compare the highest scorers against matched controls, with an assumed campaign-reach percentage in the model.
- Duplicate exclusion and suspect screening are configured separately for each eligible respondent source, and field reports carry respondent-quality histograms.
- Numeric-coded wide Excel downloads with SPSS-compatible values and an accompanying code-to-label data map.
- Cross-tab and time-series reports can include unselected options or suppress empty rows and columns. New reports include eligible options by default; existing settings are untouched.
- Customer-funded PureSpectrum accounts can be connected through the API, with provider credentials encrypted and MX8 Labs billing limited to the platform fee.
- Failed respondent source operations now return a usable error and allow correction, retry or deletion instead of stranding the source.
- Basic respondent sources can run on a single survey link rather than separate live and test links. Existing sources keep their separate links.
- A reference page for supported survey languages and markets, organised by country.
From Updates
- Catch Fraud Before You Bank the Response — Megan, on why cleaning after fieldwork is forensic work on a sample that is already contaminated.
- From Quarterly Tracker to Always-On Brand Signal — Megan, on a brand number that arrives after every decision that depended on it.
- Conjoint and HB Without the Stats Bottleneck — Tom, on the production apparatus, not the methodology, being what made conjoint a premium occasional thing.
- MX8 Labs Joins Adelaide’s AU Ecosystem for Outcome Partners — Megan, on matching survey outcomes to placement-level media quality rather than stopping at one campaign result.
- Auto-Coding Open Ends Across Every Language — Megan, on two markets answering the same way and landing in different code buckets.
- Same Budget, Ten Times the Questions — Tom, on what a research budget does once a study stops being a commitment.
Get in touch if you want a closer look at any of it, or if there is something you want next.