Research methodologies

This is the reference page for the methodologies named in the recipes. For each one you get what it measures, how it works, which question it answers and how you set it up in Exoid. Use it when you already know the decision you have to take and you’re picking the instrument to get there. Several of them can be set up in more than one way, and the setup changes what you learn. Where that’s the case, the entry says so.

​

Comparative tests

​

Monadic test

What it measures. Which version of a stimulus works better: a concept, an ad, a pack, a claim. How it works. You split respondents into equivalent groups and show each group a single version, then compare metrics between groups. Showing one version per person stops direct comparison from distorting judgement: in real life nobody sees two packs side by side knowing they’re alternatives of the same product. The question it answers. “Which version works better?” In Exoid. You assign the version with the Media field of the block: set Media source to Use variable and the stimulus follows a variable, so different respondents see different versions inside one study. The variable usually comes from the URL parameter of the link, it lands in the Variables column of Responses, and each cell becomes a comparison cluster in the same campaign. See Showing different stimuli to different respondents. Running the same stimuli as a direct comparison — Ranking or Image Choice, all versions to everyone — is the other way to build it, more sensitive to small differences and lighter on sample.

​

Pre-roll test

What it measures. The effect of a digital ad in conditions close to real viewing. How it works. It’s a comparative test built for digital advertising: the ad is shown in the environment where it actually lives, such as a pre-roll before a video, accounting for partial attention and for the option to skip. The question it answers. “Does the ad hold up where people actually meet it?” In Exoid. Load the video as the block’s media and use Wait before advancing to control the exposure time, then measure spontaneous recall, brand recall and post-exposure associations. To read the effect of the ad rather than its absolute score, add a control cell that sees no video, assigned with the media variable, and compare the two.

​

Trade-off and priorities

​

MaxDiff

What it measures. The order of importance of many items, when you can only keep a few. How it works. It solves a known problem with importance scales: ask people to rate twenty items and they’ll tell you everything matters. MaxDiff shows small subsets instead and asks each time for the most and the least important. Repeating the exercise produces a clean, comparable ranking, independent of how each person uses a scale. The question it answers. “Out of these twenty claims, which five do I keep?” In Exoid. You build the sets with two single choice blocks per subset, created quickly with Create in bulk and Add Bulk Options. Prepare the set plan first — each item appears the same number of times and with different partners — then field it, export the dataset and compute the utility scores in SPSS, Excel or your statistical tool. With seven items or fewer, the Ranking block gives you the order in a single screen and no set plan at all.

​

Choice-Based Conjoint

What it measures. How much each product attribute weighs in the choice, and how much people are willing to pay. How it works. Instead of asking how much price matters, it puts the respondent in front of complete products combining different attributes and prices and makes them choose several times. From choice behavior it derives the weight of every attribute and lets you simulate scenarios: what happens if I raise the price, add a feature, or a competitor moves. The question it answers. “At what price and in what configuration do I sell?” In Exoid. Each choice task is a single choice block between complete profiles, and Create in bulk plus Add Bulk Options make building twelve or sixteen of them quick. You prepare the experimental design with your design tool, field the tasks in Exoid, export the dataset to SPSS and estimate utilities and scenarios there. It’s the most robust method for pricing and range, and it asks for more design work and a larger sample: a considered study, not a microstudy.

​

TURF

What it measures. Which combination of items reaches the most people without overlapping. How it works. Picking the most popular items individually is often the wrong move, because they appeal to the same people. TURF looks for the combination that widens total reach instead: the second flavor in the range shouldn’t be the second most loved, it should be the one that brings new people in. The question it answers. “How many flavors do I keep in the assortment, and which ones?” In Exoid. The input is a multiple choice with no selection cap across the whole candidate range, options randomized. Export it and compute combined reach in Excel or your statistical tool. A stricter variant is Image Choice with the range shown as a shelf and a Range selection mode of one to three: you collect the basket instead of the wish list.

​

Price Sensitivity Meter

What it measures. The price range the market considers acceptable. How it works. Four questions about price: at what price does the product feel too expensive, expensive but acceptable, cheap, so cheap you’d doubt the quality. Crossing the four curves gives you the acceptable range and a few reference points. It doesn’t model choices the way conjoint does, but it’s fast and cheap. The question it answers. “What price band can I move in?” In Exoid. Four Number Input blocks in sequence, ten minutes of building. You plot the four curves in Excel from the export and read the acceptable range off the crossings. It’s the right price measurement for a microstudy. To test price as a behaviour rather than a statement, show the same product at a different price per respondent with the media variable and measure choice.

​

Brand and perception

​

Implicit association test, single

What it measures. The associations a person holds towards a brand, without asking them to state any. How it works. It comes out of neuroscience research. Brand and attribute pairs are shown and the speed with which the respondent confirms or denies the link is timed: the faster the answer, the more automatic the association. It surfaces what the consumer doesn’t know they think or wouldn’t say, and it’s most useful at the first moment of truth, when the decision at shelf takes seconds. The question it answers. “What does my brand actually evoke, beyond its stated positioning?” In Exoid. You reproduce the structure of the exercise: brand/attribute pairs one per screen, instructions to answer on instinct, Wait before advancing off so nothing slows the answer down. What you read is which associations hold and how far apart brands sit on each one. Showing a single brand per respondent with the media variable keeps the judgement free of the competitive set; showing the whole set gives you the map in one study.

​

Implicit association test, multiple

What it measures. The same associations, but across several brands at once on the same basis of comparison. How it works. Same logic as the single test. The question changes: not “how am I perceived” but “how am I perceived against competitors”. The output is a competitive map of automatic associations. In retail it compares the banner with rival banners, or the private label with manufacturer brands. The question it answers. “On which territories of meaning am I ahead, and on which am I behind?” In Exoid. The most convenient structure is a brands × attributes matrix, or one multiple choice per attribute with the brand list in random order. Transport options keeps the follow-up questions to the brands the respondent already knows, which shortens the interview and cleans the data.

​

Mental Availability

What it measures. How easily and in how many buying situations the brand comes to mind. How it works. Based on Jenni Romaniuk’s work at the Ehrenberg-Bass Institute. The logic is that growth comes more from being present in people’s heads at the right moment — the Category Entry Points — than from persuasion. You measure in how many and which contexts the brand is mentally available: an indicator of growth potential and distinctiveness. The question it answers. “In how many buying occasions do I exist?” In Exoid. One multiple choice per Category Entry Point, always with the same randomized brand list. The count of covered situations reads straight off the Summary. You can also run it the other way round — show the brand and ask which occasions fit it — and the two readings together are more informative than either alone.

​

Mental Advantage

What it measures. How strongly the brand is associated with relevant attributes compared to competitors. How it works. Also from the Ehrenberg-Bass school. Coming to mind isn’t enough: what counts is coming to mind as the best fit for certain things. It reads the brand’s competitive strength on specific territories of meaning. The question it answers. “What am I the best choice for, according to the consumer?” In Exoid. One multiple choice per attribute — “which of these brands is the most…” — with the brand list in random order. Comparison clusters show you how the advantage shifts between users and non-users.

​

Drivers and satisfaction

​

Key Driver Analysis

What it measures. How much each driver contributes to an outcome you care about. How it works. It relates a set of drivers — attributes, features, perceptions — to a business outcome such as satisfaction, brand liking or purchase intent, and estimates the weight of each. It doesn’t only tell you what the consumer thinks, but what’s worth acting on to move the needle. It’s often paired with NPS or with an equity metric. The question it answers. “What do I act on to improve the outcome?” In Exoid. You collect drivers and outcome metric in the same interview — usually a matrix plus a summary question — then export the dataset and run the regression in SPSS or your own software. Keep the drivers on one scale and the outcome on its own: that’s the shape the model expects.

​

Net Promoter Score

What it measures. Loyalty, in a number comparable over time and across companies. How it works. A single 0-to-10 recommendation question splitting people into promoters, passives and detractors. It’s more an indicator than an analysis: it tells you how you’re doing, not why. For the why it pairs with Key Driver Analysis. The question it answers. “Would customers recommend me?” In Exoid. A Rating block from 0 to 10 with Display type on Numeric, and a visibility condition triggering the open-ended question below a threshold. You set the group split with a Script or with three comparison clusters; the net score comes out of the export in one formula. Well suited to continuous tracking.

​

Penalty Reward and Kano factors

What it measures. The type of effect each feature has on satisfaction. How it works. It separates three natures. Must-have features are taken for granted: their absence penalizes, their presence doesn’t reward. Proportional ones act linearly. Delighters surprise and reward, but don’t penalize when missing. Looking at the asymmetry between presence and absence tells you where to invest: cover the hygiene factors and pick the differentiators well. The question it answers. “Which features do I develop, and which are pointless?” In Exoid. For every feature a pair of questions on the same scale — how would you feel if it were there, how would you feel if it weren’t — inside a matrix. You get the classification by crossing the two answers in Excel from the export.

​

Sample structure and depth

​

Segmentation

What it measures. How the market splits into homogeneous, distinct groups. How it works. It groups consumers so you can recognize and reach them. The need-based approach divides them by real needs and motivations, not by demographics: age describes a segment, it doesn’t define it. The output is segments with a profile and a size, usable to decide who to target and how to talk to them. In retail it becomes shopper segmentation by shopping mission. The question it answers. “Who am I talking to? Who actually chooses me?” In Exoid. You field the attitudinal battery on matrices and sliders, with a large sample and Advanced statistics mode on for the SPSS export. You run the clustering in your statistical tool, then bring the segments back into Exoid as comparison clusters, or as a typing tool with a Script writing the segment into a variable. Behavioural segmentation is the shorter route: occasions, frequency and spend instead of the battery, addressable segments and a lighter interview.

​

Multimodal qualitative analysis

What it measures. The why behind the number, in the consumer’s own voice and words. How it works. It collects open answers in video, voice and text form, with an automatic moderator that follows up and digs deeper while the person is still answering. The point is that it lives inside the same quantitative study: you have the number and the explanation from the same respondent, instead of two pieces of research to reconcile. It’s the part of the chain the market is moving on most right now. The question it answers. “Why? And how would the consumer put it?” In Exoid. It’s AI moderation on open-ended questions, plus the File upload block for photos and the Website block to collect the page people actually use. In the results you read the verbatims in the detail of a single response and explore with Talk with Data in Conversational mode. It works as a study on its own and as one extra block inside a quantitative study — the second use is the one people underrate.

​

Next steps