Insight Platforms recently hosted two full days of online demos. Each day featured seven very different research and analytics technologies: emotion analytics, communities, biometrics, knowledge management, data visualisation … a really eclectic bunch.
Each company did a 5 minute introduction; 25 minutes or so on a software walk-through; and ended with 15 minutes Q&A from the audience.
You can catch up on all these sessions here:
As I watched them live and spoke with the presenters, a few big trends started to emerge.
1. There’s some really creative stuff going on
Observation, biometrics, projective tools, video analytics … research tech is about so much more than surveys and focus groups.
Take the Maru/HUB platform. It brings together agile surveys, social media influencer analytics, implicit response modules, topic modelling of open text data … it’s kind of annoying if you run a directory of tools because it’s hard to pin it into the right categories. But don’t tell them I said that.
In a similar vein, Element Human combines several different methods to understand emotional reactions to media: direct survey questions, Implicit Association Tests, eye tracking, facial expression analysis and – if it’s available – profiling or behavioural variables from CRM records.
Crowdtech’s platform is used for big customer panels, smaller insight communities and ad hoc research. It has plenty of creative qual research tools and all the question types you’d expect for proper surveys. The video testing tool is quite neat – it’s effectively a dial test with thumbs up / down emojis and commenting, all of which is time-stamped to identify peaks or troughs in the content.
2. Good UX finally comes to research software
It’s been a long time coming.
For years we’ve had to pick our way through interfaces like Netscape from the late nineties – grey backgrounds, tiny serif fonts, fifteen clicks for every action. It’s been embarrassing.
But I have to say, some of the interfaces on modern research platforms are really gorgeous. They rival anything you’ll find in marketing tech or even consumer apps.
Check out incling, software for communities and online qual projects. The whole interface just says, “I’m really easy to use. Don’t worry, you can’t break me. This is going to be fun.”
Or Recollective, another leading player in online qual and communities. Gorgeous in a very different way, with design cues that draw from Apple and Facebook to come up with something distinctive. Clean, confident, authoritative.
Of course, looking great is only one part of the story. A pretty user interface doesn’t always make for a great user experience.
If you’re a research agency or freelancer, you need tools with a UX that helps you be more efficient. Less time faffing with setup and process, more time to spend on the valuable stuff for clients.
If you’re an insights team, you need to on-board new members quickly. You don’t want to invest two days just to hold someone’s hand as they get up to speed. Or you want your stakeholders – non-experts – to do more of their own work and relieve the load on the core insights team.
More and more platforms are designing user experiences that hit one or both of these needs.
Q One Tech makes software for panel management and surveys. It’s used by agencies and is strong in healthcare research. If you’ve ever worked in panels, you’ll know that feasibility checking is often time-consuming and fiddly – so Q One make a big deal out their simplified UX that needs a maximum of 6 clicks to check feasibility.
Sentisum is a CX Analytics platform that uses machine learning to pull insights from unstructured text. Sentisum users vary widely – they could be in teams that do web optimisation, social media, customer support, market research, e-commerce … so it’s important that any platform is simple and intuitive for non-experts. It turns out that the simplest and most effective thing is actually an automated email bulletin with an embedded dashboard.
Market Logic Software aims to be the Insights Platform for large multinationals. It is a planning / management tool for insights projects; a research repository / knowledge management system; and a self-service platform for business users (brand managers, media planners, senior execs).
Everyone knows these people have no patience and never read a manual, so good UX is critical to drive adoption and avoid stakeholders phoning the insights team with every little question.
3. Survey data is weird and needs expert tools
It’s a fact. The way survey data works – multiple response formats, skip patterns, loops, order effects, missing data – is annoying. It’s nothing like the data that comes from website analytics, transactional systems or retail panels.
Every time I’ve consulted with companies doing top-down digital transformation, this has been Brand New Information for their IT, BI or digital teams.
“We’re going to build a giant lake with all your survey data and all the other customer data. You’ll use [Looker / Tableau / Whatever]. And the sun will always shine.”
The Demo Days featured a few companies who know what I’m talking about – and have built software tools that handle survey data properly.
Infotools Harmoni is a data analytics and visualisation platform that loves big, ugly surveys – like 63-country tracking studies for Shell. And – unlike most data lake / BI platform combos – actually can integrate survey, transactional and behavioural data in one place.
If you want a more articulate view on the limits of BI platforms for survey data, watch this excellent webinar by Sjoerd Koornstra, former insights leader with Heineken.
Askia also understands the nuances and challenges of survey data. It mainly serves research agencies with its suite of survey, analytics and visualisation tools. Like Infotools Harmoni, survey data is its bread and butter. Plenty of messy, multi-market brand trackers also run on Askia.
And surveys are not just about the data from closed questions. Open-ended questions can generate all sorts of headaches for agencies and insight teams. For a long time, the only way to get structured meaning from verbatim answers was to have it all manually coded.
Costly and slowish.
Then came text analytics.
Cheaper and faster. But too often, generated garbage, owing to small training data samples.
Codeit was developed to try and square this circle – to make manual coding faster and cheaper with judicious use of some machine learning and some smart workflow tools. Agencies who use it code more open ended responses – and manage to spend less on coding overall.
4. Qual and UX research need better tools than Zoom
Like most of us, after months of lockdown and remote-only meetings, I have a love/ hate relationship with Zoom.
I love the fact that it just works.
I hate the way it adds bloat and shine to my face. Oh well.
Anyway, there has been a stampede recently from offline focus groups to online video discussions using Zoom, Teams, Google Meet and whatever else is to hand.
For qualitative research, it’s a depressing version of Ford’s faster horse. It’s like the dawn of online surveys, when agencies copied over 30-minute CATI scripts to torture early internet users.
Thank God nobody runs 30-minute online surveys these days.
My point is not that Zoom groups are inherently bad. It’s just that they really shouldn’t be the default for online qual. We have so many better options that make fuller use of the digital medium.
If an online video group really is what you need, FlexMR InsightHub gives you the tools to run it properly – privacy controls, observer rooms, options to use text chat alongside or as an alternative, stimulus sharing and markup. It’s built by / for researchers rather than as a generic business tool.
But all four of these qual-focused platforms also let you run text chat discussions, asynchronous groups over a few days, short and long term communities, creative brainstorming exercises, whiteboards, stimulus markup, mobile ethnography, diaries, screen capture for UX feedback … I could go on.
There really is so much more to online qual than video groups.
If you want to learn more about creative online qual methods, watch this free webinar from Tom Woodnutt.
Or join his three-part online training course in September. People who joined it in May said it was a belter.
5. User research and market research are inching closer
OK, I might have said this before …
But if you read that article, one of the main ways I see the disciplines converging is in the use of technology platforms.
Most of the qualitative platforms are used interchangeably for user and market research: for interviews, documenting journeys, persona-building, in-situ observations. Even narrated screen recordings – which are widely used for ‘think aloud’ usability feedback – as well as market research applications like online path-to-purchase mapping.
6. AI is everywhere
By the end of this year, ‘Artificial Intelligence’ will sound a bit quaint and old hat.
It will start to disappear from our lexicon – like all technology labels do once the underlying tech is everywhere.
When technology is widely spread, it becomes effectively invisible.
This is what’s now happening with AI features in research and analytics platforms.
Heartbeat Ai uses Natural Language Processing to model the emotional content of language from similar sources. It can use text-speech algorithms to convert audio messages and extract the emotional content in near real-time.
Sentisum applies similar techniques to get meaning from unstructured text data – in surveys, complaints, reviews, comments, social posts – and make it useful for different teams.
Codeit uses some of the same principles to speed up the coding of survey open-ends.
Element Human uses Computer Vision to convert pixel-level changes – eye movement, facial muscle movements – into emotional signals.
Market Logic Software uses AI to build its knowledge graph (millions of semantic connections between insights documents), make recommendations based on prior searches and proactively suggest answers to questions.
Q One Tech uses AI to get sample feasibility results in its panel management software.
See what I mean? ‘Artificial intelligence’ is everywhere right now – qual, quant, surveys, analytics – which surely means it will soon be nowhere at all as we stop using Star Trek language to describe useful everyday features.
BTW don’t tell anyone that AI will soon be old hat.
At least not until we’ve run the AI Summit in September.
7. Video is everywhere
More than half of the platforms who showcased their wares have video as a core part of the offer: video groups, interviews, clips, vox pops and enthographic nuggets (FlexMR, incling, HatchTank, Recollective); video archive, search and showreel creation (Market Logic, with integrations); video testing in various formats (Crowdtech, Element Human, Maru/HUB).
I guess this is a logical extension of the ‘Ai is everywhere’ thing. AI makes it feasible to analyse, transcribe, search and clip video for a reasonable budget.
8. The future is better integrated
APIs are the Babelfish of the internet. They translate between different software apps so that data can be exchanged nicely and safely. I couldn’t run my business without Zapier, which ties together at least a dozen pieces of my own tech stack. So it’s heartening to see how much focus these platforms are placing on integration.
Codeit has direct integrations with Decipher, SurveyGizmo and other survey tools – so that coded verbatim data can hook straight back into the survey database. It has also been an early partner helping to develop TSAPI, which aims to be a new common standard for interchanging survey data between systems.
Remember that monster tracking survey in Infotools Harmoni? It connects to 13 other sources as well as the survey data.
Element Human’s HX Embed product allows pretty any survey platform to include facial biometrics as a component.
And Market Logic integrates with dozens of third party data sources and software tools.
It’s so nice when we all get along.
9. But it’s not all about The Revolution
“People don’t want something truly new, they want the familiar done differently.“
Not my words. That was Nir Eyal, author of Hooked (essential reading if you want to know why you check your phone five hundred times a day).
If what you’ve read here so far sounds like too much new stuff, try not to worry. Most of the energy from these tech firms is focused on making existing research processes better rather than building something radically different.
Codeit wants to improve how open-ended responses are handled. Askia wants to make it easier to work with tabulated survey data and charts. Market Logic wants us to use all the research data we already have. Hatchtank wants to make it easier to work with qual video.
Surveys, focus groups, interviews – none of this is going away any time soon.
Yes, there are some more revolutionary features and ideas coming from these platforms and others – but it’s much more about working within existing frameworks, not tearing down the house and starting over.
Watch all the sessions from the July 2020 Demo Days here, and look out for the next sessions which will run in October 2020.