Select a type to quick search

Saros is an established qualitative and UX research participant recruitment service, with a bespoke database of over 300000 UK consumers - NOT a quant panel, they're genuine individuals, keen to take part in qualitative research.

We have access to a constant supply of new people to take part in your research through investment in building our database.

Saros combines the latest technology with high quality screening and excellence in project management - to deliver participant recruitment that meets both hard and soft criteria. From depths to groups to online communities and UX interviews, we can find the right people.

Charities/Voluntary, Drinks (Alcoholic), Drinks (Non-alcoholic), FMCG – General, Food, Gambling/Online Gaming, Healthcare, Home Entertainment, Home/Garden/DIY, Media (Broadcast), Media (Mobile), Media (Print), Online, Pets/Petcare, Pharmaceutical, Retail, Sponsorship, Sport/Leisure/Arts, Telecommunications, Toiletries/Beauty Products, Toys/Games, Transportation, Travel/Tourism, Wellness/Fitness
Creative development research, Depth Interviews, Diary Studies, Ethnography, Face-to-Face, Group Discussions/Focus Groups, In-Store Interviews, Online Communities, Qualitative, Recruiting, Web Usability Research
Consumer, Crowdsourcing, Customer Communities, New Product Development, Online Panels, Packaging/Design, Product Testing, Social Media, Social Research, Usage & Attitude
Affluent, Baby Boomers, College/University Students, Digital Consumers, Kids, Mature/Midlife, Mothers/Parents, Senior Citizens, Women, Youth/Teens
Senior Contacts

Shelley Hodgson (Associate Director)
Fiona Jack (Managing Director)

Breakdown of Personnel

Admin/Support staff: 2
Executive/Research staff: 7
Telephone interviewers: 8
Total Number of Employees: 11 to 20


PO Box 71506
SE10 1BX
Tel: +44 (0)2084817167
Establishment date: 2000

Recruiting qualitative participants research using quantitative algorithms

Qualitative and quantitative research are very different disciplines. As everyone knows, qual is never intended to be statistically correct or representative of any demographic segment, instead its role is to explore in depth issues and feelings in relation to a defined area of interest under consideration.   

This couldn’t be more different from statistical research… however when it comes to recruitment for qualitative research, it is not unusual to find a desire expressed to use quantitative tools as part of the selection or filtering process.

One regular incidence of this is the use of bespoke algorithms to define segments or personas. Many of these are proprietary tools, which have been developed in secret by brands and their partner agencies often over many years – they know their customers from extensive quantitative profiling, and desire to model them in order to predict future behaviour.  These can indeed be used in qualitative recruitment, but the ease of doing so depends greatly on the complexity and secrecy of the algorithm. Often we are handed what are effectively black-box tool, usually in the form of a spreadsheet with a locked macro.  It consists of a series of questions, for which you input the answers into the spreadsheet, and a result it is spat out:  “The person providing these responses is a typical fit for the "yellow" profile”

There are many issues for us doing recruitment using it to like this. That's not to say it can't be done, but it can be more difficult than first expected.

Part of the issue is with the secrecy, when we have no idea what defines a yellow persona as opposed to a red blue or green one. Of course we would love to sign your client's NDA and be told a little bit more about it, if we possibly can. This might seem an unusual request from your recruitment partner but it is exceptionally valuable simply as a sense check – our outlook on the process is qualitative after all. If blue segment users are supposed to be late technology adopters, innately conservative , family life-stage people – we might find it valuable to know this and compare against other answers given in recruitment, and what else we know about that applicant based on their previous projects and their membership profile on the Saros database. Because algorithms are not foolproof, people are not personas, and it is our job to get a rounded qualitative sense of the individual's likely fit in relation to their profile for which we are recruiting.

This is particularly important because very often we know these algorithms are worked out based on data from vast quantitative sources. This data must have contained atypical responses and margins of error – it's a tendency towards a persona, rather than a description of a unique individual. If we were recruiting you a panel of 2000 people and we ran them all through the algorithm, there is a good chance we would end up with a panel that contained a lot of people who strongly fit the algorithm and embodied the persona. The panel would also contain extremes who were less of a fit in some ways than others, but still scraped through the scoring process – probably clustered in the kind of bell curve that the stattos would love.  Work you did with that panel would over time balance out any discrepancies provided by these outliers who less typified the persona – simply because of the scale it involved.

However in qualitative research we might be recruiting you a single focus group of eight people.  It is essential therefore that we know a little bit more about what is going on "under the hood" of the algorithm tool, that we understand the thinking behind using this too as part of the qualitative selection process. Because otherwise if we come across an outlier who nevertheless screens in, there is every chance we will end up recruiting somebody who is perhaps completely wrong for your project – and we won't even know why.

There are often practical issues associated with applying algorithms in use as well, not least because they frequently involve large numbers of questions. If we can be provided with the right to unlock the algorithm and see the coding behind it, we may be able to incorporate it into our online screener – use it as part of our pre-filtering, and make sure that we are in fact only calling people who type out strongly as a yellow. This means that we can then focus during our interview on the description of the yellow persona, the softer intangible qualities, and making sure that we recruit people who are the best possible candidates for your research.

However if we are provided a locked tool, the only thing we can do is apply it after the event during the telephone interview stage – this is more cumbersome and expensive, because it does not enable us to rule out people who are not a fit before the calling stage.  Depending on the expected incidence of the desired segment(s), the strike rate – and therefore costs involved in recruitment – may even approach that of cold-calling. That is often a surprise to clients, but it is a consequence of trying to use quantitative tools in qualitative research. 

Excerpt from “The Participant Principle”, UPP Books 2016

Source: LinkedIn
Katarzyna Kotecka-RavasiSenior UX Research Coordinator/Recruiter at Google GmbH

"The Saros team have been recruiting participants for our usability studies for several years now and we have always been very satisfied with the service provided by them. Their professionalism, commitment, flexibility in accommodating last minute requests as well as high quality of participants provided, make them a great partner to work with"


Barbara Langer Head of Market Insights, eBay Europe

"Saros is one of the most flexible, accurate and excellent value market research recruitment agencies that I've ever worked with. Never thrown off by last-minute changes to recruitment profiles, always keen for feedback, and just a supreme pleasure to do business with." 


Get the latest MRS news

Our newsletters cover the latest MRS events, policy updates and research news.