Understanding People

A daily measure of public perception of more than 1,000 consumer brands across 40 sectors, measured on a 7-point profile, with data delivered next-day. We interview 5,000 people each weekday, more than 1.2 million interviews per year. Respondents are drawn from our online panel of more than 1,000,000 individuals.

Who is BrandIndex for?

All those who need to know what consumers really think of a brand - and its competitors.

For the companies in our BrandIndex survey, the CEO, who is the person most responsible for brand equity. Second, the CEO's executive team – Finance, Marketing, Communications, Human Resources, as well as brand managers and the research department.

BrandIndex is also valuable to advisors and analysts of the BrandIndex companies. Public Relations firms and Advertising Agencies can use BrandIndex to monitor the effectiveness of their campaigns, and share this data with their clients. Financial analysts will find BrandIndex information valuable for its consumer insights, which are critical for the future prospects of any company.

Why do I need continuous data?

Most brands do not shift very much from month to month, let alone day to day. However, all trends begin somewhere. And with BrandIndex, you can spot those trends early, and take advantage of an opportunity or defend against a potential threat, before it is too late. A competitor might be showing weakness in its loyalty or satisfaction scores, presenting a opportunity for you gain market share. Or a competitor may be closing the gap with you on Quality or Value, prompting you to take defensive action.

And even though most brands don't change very much in the short term, most brands will experience 'turbulence' from market events, and managers will benefit from having an immediate reading on the extent and nature of any effects. Though your own brand may be stable – what about your competitors? In any sector, things will be happening every single day: a new product is launched, an advertising campaign begins, a PR disaster occurs – maybe a strategic PR campaign to change a brand's image begins. Whatever it is, you will want to know exactly what effect that's happening on your own and on your competitor brands.

How easily can I access the data?

A series of pull-down menus produces graphs for you within seconds. It couldn't be easier. The whole point of this product is to take high-quality, high-relevance data and put it straight into the hands of decision-makers. So the reporting tool has been designed to suit the style of decision-makers: it's instant, it's clear, it's easy – and above all, it makes sense of the data.

How does this product compete with other similar trackers?

Some companies have existing, customized, brand tracking capabilities. In these situations, BrandIndex will be a great addition to their current tracker, enhancing their ability to get an early read on changes to key brand health indicators, and then either take immediate action, or perform more in depth analysis using their custom systems.
In other cases, companies may decide that BrandIndex presents an opportunity to rethink their brand tracking strategy. BrandIndex can provide the high level tracking, and also offers the opportunity to do more in depth research with our panel members, when the data indicates that additional analysis would be beneficial.

What does the data actually mean? Explain the difference between the various scores.

The score is the 'net rating': people are asked to identify the brands to which they have a positive response, and then those to which they have a negative response, to whatever is the prompt measure. The net score is the positive minus the negative.
The charts also have bar-charts along the bottom. These represent the number of people who made either a positive or a negative identification – we call it the 'mindshare', the number of people who have given a view of the brand. A brand given a positive rating of 15%, and a negative rating of 10%, would have a net score of +5% and a 'mindshare' of 25%.
The seven measures that make the complete profile are below. PLEASE NOTE that each measure is taken independently – in other words, in any one survey, any individual respondent is asked about only one measure for the sector, not all seven. Therefore none of the readings influence each other within the survey. We attach great importance to this is making the data more scientifically robust.

'Buzz': have people heard anything positive or negative recently about any of the brands in a particular sector? Respondents are prompted for anything they've heard in the media – news or advertising – or in conversations among friends and family. This is to elicit whether people have noticed good or bad news, advertising or PR campaigns, product launches - or whether there is any 'word on the street' emerging.
General impression: for which brands in a sector do people have a 'generally positive' or 'generally negative' feeling?
Quality: which of the brands in a sector do people associate with good or poor quality? (We remind in the prompt that this is about the quality of the product or service irrespective of its price).
Value: which of the brands do they associate with good or poor value-for-money? (We remind in the prompt that this is about 'how much you get for your money' and not about the quality).
Satisfaction: would the respondent identify himself or herself as a recent satisfied or a unsatisfied customer of any of these brands? We leave 'recent' undefined as it has different meanings for different sectors, and respondents will adapt it to their own experience.
Recommend: which would respondents recommend to a friend? Or suggest they avoid? Shifts in this figure could suggest future change in market share.
Corporate reputation: we get at this through a proxy: which of the brands in this sector would people be proud – or embarrassed – to work for? We prompt that this question is to be considered in terms of working for the brand or business, in the kind of job they do at the moment (or used to do) so that, for example, the association isn't necessarily with flipping hamburgers for a fast food outlet, but the reputation of the corporate owner of the brand.

In addition, we supply an Index score, as follows:

INDEX: the average of measures (2) – (7) above. 'Buzz' is excluded as this may (or may not) be an effect on the brand but not yet part of the underlying perception. One of the things we want to track is the effect of Buzz on the brand profile.
INDEX-R: this is the same as the INDEX score except that it is adjusted relative to the sector average. In a league table of brands, tobacco and fast food brands tend to crowd out the bottom, because there is quite strong disapproval of these sectors. Nevertheless they also enjoy huge consumer support. We think a brand should also be judged on how it does its job as a brand – that is, to distinguish itself from other competing products and services. The ability of a particular brand to outshine its sector rivals is as important to know as the degree to which we have an overall positive or negative view of it.

Why is the data reliable?

Establishing the accuracy of data of this kind is always difficult. However, YouGovPolimetrix is in the unique position of having demonstrated its accuracy in the one area that it can be properly evidenced: in predicting election outcomes. In 2006, the Cooperative Congressional Election Study, sponsored by 37 universities, analyzed 40 senate and governor forecasts released prior to the elections. YouGovPolimetrix outperformed all other Internet and telephone polling companies in predicting the correct outcomes.

Research has also shown that online research gets closer to the real views of people than conventional research which uses human interviewers. In one study, people were asked which of a list of bottled water brands they recognized. They were asked by telephone, face-to-face, and online. One of the brands was non-existent – but it scored second-highest in the face-to-face and telephone interviews (with 24% and 22% recognition). Only 2% of the online sample claimed to recognize the non-existent brand. There are many other studies that back up this advantage of online research.

Another advantage is that respondents answer at a time of their choosing and convenience, and see the questions in a format (on-screen) in which they are used to getting information. So this is a natural and sensitive setting for conducting research, producing better data.

Also, we are asking large number of people. While the daily figures are not huge samples, the accumulation of views day after day means that large samples can be examined and analyzed in detail – for example to explore and compare the views of various demographic groups.

How big is the sample? What is the sample design?

We interview 5,000 people each day. The sample contains representatives from all relevant demographic groups, but for this product we have over-represented higher income and more highly educated individuals, as these tend to be more interesting to those tracking brands.

What is a "day" in terms of the data? And when is it reported?

A ‘day’ is the 24 hour period from midnight to midnight. We collect data Monday through Friday. Respondent data is processed and uploaded into the reporting tool by 9am every business day morning.

What is the survey experience for the respondents?

Don't they get tired of all these ratings?

How often do they answer the same question?

How long is the survey for them?

Why would they take part?

Our panel is proactively recruited, mainly through surveys on topics that interest them on their own favorite Internet sites, or when conducting Internet searches. We invite them to our panel, and have a modest incentive system, in which they are awarded ‘points’ for every survey that is completed. Over time, they can redeem points for modest prizes such as t-shirts, tote bags, and gift cards.

The BrandIndex survey has been designed and tested to ensure that respondents have a generally positive experience. We begin and end each survey with questions about their lives, about things happening in the world, so that our panelists have a varied and interesting experience.

They know they are contributing to data not only about consumer issues, but about the wider world. This is important to them, and improves engagement with the activity - which in turn ensures a higher quality of data.

The typical panel member will take this survey four times per year. The maximum anyone can take the survey is once per month. But please note that the structure of the survey means that a respondent will only rate a particular sector on a particular aspect once every 14 surveys - that is, less than once per year. Also note that regular engagement with a survey design means that responses become more - not less - reliable, as those taking the survey understand better how to report on their experience and thoughts (so long, of course, that it is not too often!).

What's different about the survey design?

We have used the flexibility of online research to come up with a new design for BrandIndex. It is a design which we know is convenient and responsive to those taking the survey, and which produces robust and sensitive data. Instead of asking people to rate each brand separately - a mind-numbing task which would lead to very poor data - we present them with a screen of brands in a particular sector and ask them to click on all those which (for example) they associate with value-for-money. And then we ask them to review the same brands a second time, to click on those which they associate with poor value-for-money.

When a respondent has answered on one aspect on a sector (e.g. value-for-money), he or she will not rate that sector on any other aspect. In this way - and this is important - all measures are independent of the others. Differences in the profiles are therefore reliable.

What is the survey design?

There are forty different routes through each day's BrandIndex survey, with each respondents being asked first to rate aspects of their own lives - the health, happiness, work, their sense of what is happening in the neighborhood, the country, and the world. This data is used for other tracking purposes, and establishes a sense of involvement on the part of the respondents.

The survey then presents a screen of brands from one sector, and asks the respondent to click on those brands which are relevant to the prompt question (e.g., 'of which do you have a generally positive impression?') The same sectors / brands then appear with the negative form of the question. This happens six more times, each time with different sectors and different questions, until the survey ends with a series of ratings about things in the news or related subjects. (This ending is important, because it 'book-ends' the consumer research with personal, political and social issues, which we know increases respondent interest and engagement, and therefore the quality of the data. It is one of the things that differentiates being on a YouGovPolimetrix panel from being on other consumer panels).

We take great care that the respondent experience is generally a positive one, and that they will be willing to take the survey again when invited. After each BrandIndex survey, we ask for feedback from respondents on the survey experience, in an effort to continually improve the survey.

How were the brands chosen?

Not all brands can be included – in any sector, there is a very long 'tail'. We worked with beta clients to determine the initial list of brands and sectors. Over time, we have made modest adjustments to the brand list, and will continue to modify the lists to better meet emerging market and client needs.

Why did you did not include all brands sectors?

And why did you define the sectors as you did?

This is already the most ambitious research project of its kind. We need to set some limits, and there are some sectors that we did not include. We may include them in the future. Nor did we want to use all the subdivisions sometimes used by the market research industry. Our focus is on the real world of the consumer and our clients.

When did data collection start?

YouGovPolimetrix has been collecting and testing sectors and brand data since early 2007. A subscriber to BrandIndex receives full sector and brand historical history dating back to June 1, 2007.

What exactly do I get for my subscription?

You receive the entire BrandIndex service, which includes all of the brands across all of the sectors, all of the demographic breaks, and all updated – daily. In addition, you receive the historical data going back to June 2007. And there is no per seat user charge for access to the system. We want you to be able to place the BrandIndex information into the hands of the relevant personnel, without hindrance. YouGovPolimetrix also provides initial training and updates on new enhancements.

Are there any extra costs?

No.

How is it that a big, valuable brand like McDonald's gets such a low score?

Some brands have many satisfied customers, but also generate some controversy. Fast food outlets and tobacco companies are examples. Those who dislike the brand, but were, perhaps, never likely to be customers anyway, mask the success of these brands. The key issue is change - whether the numbers for McDonald's go up or down or stay steady.

Of course, a BrandIndex subscriber can also compare its scores to the appropriate sector average, to obtain a measure of how it performs vs. its peer group. So a brand which outshines its competitors in a less popular sector can be evaluated appropriately.

Why did you use brand names, instead of the brand logos, in the BrandIndex survey?

We tested the difference between using brand names and logos in a mixed-sector list. We found in most cases, it made no significant difference whether we asked using the name or the logo. Where it did make a difference, that difference was removed when we had a clear sector context for the brand name. We therefore decided to use the name rather than the logo, as that would enable us to track over a long time-period without concern about changing designs.

Why are you not publishing the daily data in conventional tables?

We want to make this a dynamic product which the non-specialist, as well as the specialist researcher, can use. Market research is too often made inaccessible to decision-makers: we want the research to be at the finger-tips of the CEO as well as the brand managers.

Why is Internet-based market research better than the old-fashioned methods?

It is not necessarily better in all circumstances. There are many poor-quality Internet-based research companies as well as good ones, just as there is mixed quality in the conventional sector. Nor is Internet-based market research always an appropriate method. One of the disadvantages of most Internet-based research is that it is often unable to reach the poorer segments of the population (just as conventional research has difficulty reaching particular segments of society). In the case of YouGovPolimetrix, our unique proactive panel recruitment methods mean that we can supply online nationally-representative samples on a daily basis, which makes it highly appropriate for this kind of research.

Once the quality of the sample is assured, the online methodology has obvious advantages. It is quicker and more cost-effective, for example – a product such as BrandIndex would be very difficult to provide at an affordable price using conventional research. But more importantly, online research (we believe) is better at getting reliable data. Respondents complete the survey when they want to, not when a pollster disturbs them out of the blue. And they receive the questions in a format in which they typically digest information – onscreen. There is no hurry. There is no human influence to make them feel they should say one thing or another.

The accuracy of YouGovPolimetrix research has been confirmed in the only way possible – by comparing predictions with hard outcomes (i.e. elections). In 2006, the Cooperative Congressional Election Study, sponsored by 37 universities, analyzed 40 senate and governor forecasts released prior to the elections. YouGovPolimetrix outperformed all other Internet and telephone polling companies in predicting the correct outcomes.

Since when has YouGovPolimetrix focused on consumer research?

YouGovPolimetrix is well known for its political and social polling, published often in the media, especially in The Economist. We are also working with CBS News throughout the 2008 presidential election. YouGovPolimetrix has conducted market research ever since it was founded in 2004, and market research is expected to make up the majority of its revenue in 2008 and beyond.

What does YouGov mean?

The name 'YouGov' was intended to reflect the new world of consumer power: you are in charge, not only democratically but as a shopper!

What kind of customer support can I expect?

BrandIndex is designed to be easy-to-use. However, if you do have problems – for example setting up your daily profiles – we will provide the personal support you need to solve the problem. We also have a customer help line and online help facility.

Are there any other BrandIndex-related research products from YouGov?

In addition to the syndicated BrandIndex service, subscription clients have the ability to commission custom research projects and surveys, based on our BrandIndex panelists. This follow up research can provide valuable insights into the causes behind the BrandIndex scores, and the changes in those scores over time.

Is BrandIndex available for other countries?

BrandIndex is currently available in the U.S., United Kingdom, Scandinavia, Germany and the Middle East. We expect to launch BrandIndex in additional geographies soon.

How do new brands get on?

Each quarter, YouGov will review the brands that are included in the BrandIndex. Brands that become important to the sector will move in - others may drop out. Clients interested in a particular brand can pay to have it included.

What are the moving averages? Which averages are robust?

We provide you the flexibility to select from a variety of moving averages. The appropriate moving average will depend primarily on the number of respondents in the selection set:

For each individual indicator for each brand, there are 125 daily respondents.
The Index is based on the average of six individual indicators (General Impression, Quality, Value, Corporate Reputation, Satisfaction and Willingness to Recommend) there are 750 (6 indicators * 125 respondents for each indicator) daily responses for that measure.
As you introduce demographic breaks, the daily response set can be reduced significantly.

Therefore, as a general rule:

3, 5, 7, and 10-day moving averages are typically appropriate for looking at Index scores.
10, 14 and 20-day moving average are typically appropriate for analyzing individual indicator scores.
20 and 30-day moving averages are typically appropriate for analyzing long-term indicator trends, or for viewing data using demographic selections. The more demographic selections, the lower the response set, so the larger the moving average would need to be.
In times of extreme volatility due to a major event (e.g., an airline grounding hundreds of flights due to safety issues, a food company announcing a recall due to contamination) it may be appropriate to analyze 1-day scores.

Indicative margins of error for the different sample sizes are given below:

315 (demographic break on one-week single-measure score): +/- 5.5%
625 (one-week single-measure score): +/- 4.1%
325 (demographic break on daily INDEX score or one-week single-measure score) = +/- 5.4
1,500 (demographic break on monthly single-measure score): +/- 2.5%
3,000 (monthly single-measure score): +/- 1.8%
3,250 (weekly INDEX score): +/- 1.7
15,000 (monthly INDEX score): +/- 0.8

Which demographic segments can I look at separately?

The online reporting tool allows you to break the data on the following demographics:

Gender
Age
Ethnicity
Education
Income
Geography

Can I set it up so that I can easily check the same brands each day?

Yes, you can set up a series of profiles so that at a touch of a button you will see all the brands or aspects of brands that you like to check regularly.

How much does it cost? How do I subscribe?

Please contact us at brandindex@polimetrix.com

How is the PollingPoint panel recruited? What type of validation is conducted?

The goal of our recruitment is to create a large and diverse panel, which enables to accurately survey the U.S. population, and virtually any subset of the U.S. population. Our proprietary online panel community is branded PollingPoint. YouGovPolimetrix employs a double opt-in process to recruit and confirm all panelists. We typically create recruitment campaigns targeting specific demographic groups, and then leverage Google AdWords to position these recruitment surveys alongside relevant search results (for example, a recruitment survey designed to recruit men might be focused on which team will win the NBA playoffs and that campaign would appear in the advertisements alongside a Google search focused on the NBA or the NBA playoffs). At the end of the recruitment survey, a potential panelist is invited to join our panel. If the potential panelist responds affirmatively to the request. YouGovPolimetrix then sends out a confirmatory email to the respondent. Only if the potential panelist opts in to this email, does he/she become a panelist. Also at this time, we collect address and basic demographic information. Going forward, as the panelist takes surveys, we continue to collect additional attitudinal and behavioral information, to build out our ‘profile data library.’

YouGovPolimetrix is a leader in online sample quality. Using a proprietary and patent-pending technique called sample matching, panelists are selected from the PollingPoint panel, aligning those invited to participate in surveys with a compendium of national, state and local voter lists as well as other consumer databases across a number of demographic, geographic and socio-economic linkage points. In addition to providing representative samples, YouGovPolimetrix regularly cleans its panel database using sophisticated algorithms to identify fraudulent panelists, panelists who regularly skip responses, panelists who straight-line, and panelists who provide inaccurate data in other ways. We also cookie our panelists to identify fraud and we check answers from previous survey to identify inconsistency.

In addition to the standard demographic breaks, are further demographic breakdowns possible?

Yes. We collect a variety of demographic information on our panelists, including birth dates, income levels, zip codes, etc. Our current demographic breaks in BrandIndex conform to industry standards. When conducting re-contact or custom research against the panel, clients are able to customize the demographic breaks.

What if my footprint is regional, rather than national? How is BrandIndex relevant to my brand or business?

Our sample is intended to be representative of the U.S. population. This includes our geographic breaks. So you can focus on the region(s) that matter most to you, and be assured that you are looking at respondents who are representative of the geography(ies). For custom surveys, we can even identify respondents by relevant zip codes.

FAQs | BrandIndex Case Studies | BrandIndex Newsletter

Contact Information:
T: 646.723.4646
F: 212.226.7186
E: brandindex@polimetrix.com