DESIGN PHASE

ITC508 - 201990 Assignments Case Study, v1.0

Fynapa – Your partner for market research

Fynapa is a global digital data collection company, providing market research insights to its clients.

Their information system is quite complex and needs a complete overhaul. Following paragraphs

explain their business operations and as such the information system requirements can be derived for

this narrative.

Fynapa primarily provides market research services to their clients. Their clients are either individual

researchers conducting research involving data collection through surveys, or businesses who need to

gather insights about their products, services, marketing and perceptions. Survey participants are

called panellists, and anyone can register as a panellist — they get paid to participate in surveys.

Fynapa boasts a globally distributed and extremely diversified panel containing people from all sorts

of demographics and professions.

To register as a panellist, users have to sign up at Fynapa consumer website by providing some basic

profile information including an email address, gender, date of birth and residential address. Once

registered, they can proceed to provide lot of optional information about themselves, for example,

their demographics, interests, hobbies, profession, industry, job title etc. The more information they

provide, the more chances they get to participate in surveys relevant to them. Some panellists also opt

to provide their legal identity documents to become verified members. Verified members can

potentially make higher earnings. Fynapa forwards the ID documents to an external ID verification

service.

Researchers or businesses must register and login to Fynapa client website in order to create and

conduct surveys. A client account can be registered with just an email address. Once registered, clients

can go to create a new survey. According to business rules and the area of research, each survey has a

target panellist audience. It can be illustrated by a few examples: A confectionaries business in

Australia wants to know how their new chocolate advertisement is being perceived by parents of

toddlers. A social scientist wants to explore the video game addiction levels in teenagers belonging to

wealthy families. A holiday-tour organizer is interested to know what destinations are currently

popular among frequent travellers. Therefore, the clients need to specify some conditions to limit

survey availability according to panellist attributes and habits (age, ethnicity, financial capacity, family

details, likes and dislikes etc.). A user-friendly interface helps clients/researchers to define such

participant restrictions.

To design a new survey, clients are presented with another friendly interface. Surveys can be

composed of several different types of questions, e.g. multiple choice, drop down selection, multiple

check-box selection, how likely you are to recommend something (1 to 10), agreement scale (1 to 10),

match columns, numeric response, textual response etc. For each type of question, an appropriate

interface is available to facilitate creating the questions. While a survey is being prepared, clients have

the option to save it as draft. Once a survey is designed, there is an option to run it in preview mode.

After finalizing the survey, clients mark it as ready. Next, they proceed to make an advance payment

for the survey. Upon completion of payment, the survey goes live.

Fynapa’s pricing methodology is a bit involved but fair. Firstly, there is a base survey listing cost that is

applicable no matter how long the survey is. Secondly, there is a cost to conduct the survey that

depends on survey length, type of questions and the number of responses needed. Every type of

question has its fixed cost, hence the total cost is calculated by adding up individual question costs.

This base cost is then multiplied by 100, the minimum number of participants for each survey. Cost per

survey reduces after the first 100 participants. Additional responses are charged at a gradually

decreasing rate. Clients have to choose a required number of responses after which the survey closes.

For example, a survey consists of 5 multiple choice questions, worth 20 cents each, and two short

answer questions, worth 50 cents each. Thus the base cost is $2. In addition to the fixed survey listing

cost, in this example, the client need to pay $200 for at least 100 survey attempts. Let us say the client

has a target of 600 surveys, the cost of every subsequent batch of 100 is slightly smaller, such as $195,

$190, $185,….. down to some predefined minimum. Before a survey goes live, clients have to pay an

advance payment which includes the listing cost plus half of survey conduction cost. The balance is

payable at the end of survey before the results are released to clients.

The number of participants required for each survey is defined by client business requirements. There

is also an expiry date for survey availability, once again defined by the business/research requirements.

Some clients may want to survey ID-verified panellists only in order to receive more accurate results

at a higher expense. Clients are notified by email when the required number of responses are received

for a survey. Depending on their budget, clients may opt to keep the survey running to receive more

responses.

Panellists get notified of new surveys by email. Clicking the link in email will lead to the survey only if

the survey is still running, otherwise the links expire. The email also contains information about the

survey topic, expected duration of survey and the monetary reward participants will earn if they

manage to complete it. Based on participant restrictions as defined by clients, all panellists with a

matching profile and habits are invited. Additionally, panellists with incomplete profile information are

also invited, but upon starting the survey, they are screened via some auto-generated eligibility

checking questions. These questions look like the following: Do you live in Perth, WA? How much is

your monthly earning (choose from five groups)? How often do you consume biscuits (once a day, few

times a week, rarely) etc. If the panellist is eligible, actual survey starts otherwise customer is shown a

sorry message. Such screening questions also help fill the gaps in panellist’s incomplete profile.

Most surveys are designed to be completed in one sitting. But some long surveys are meant to be

completed over a longer period, for example doing a diary every day for a week, or recording the

behaviour once a week for two months. For these kind of surveys, once the survey is started,

participants are sent a unique link in email to resume it later on. If the link is lost, all the survey progress

gets lost as well.

Upon successfully completing a survey, panellist account is credited with the corresponding reward.

Partially completed surveys are not eligible for any reward. Before the panellists can redeem their

rewards, they need to earn a minimum threshold amount, which varies by the country of residence.

They also need to provide their mobile phone number as a verification of their location. Available

redemption methods can include bank cheque, direct credit to bank account, PayPal or gift vouchers

— which methods are available also depends on country of payout e.g. Gift vouches are not available

for residents of Singapore. All payment methods also have their own minimum requirements, e.g.

account deposits in Malaysia require a minimum transaction of 15 MYR. A user may also opt to donate

their credit towards one of the available charity organizations.

A very crucial business requirement for Fynapa is to nicely present the survey results to clients — this

is how Fynapa plans to differentiate itself from its competitors. A variety of reports are prepared for

each survey. The most type of reports are, of course, the survey responses. Some question results

(multiple choice, numerical response, agreement scale) can be visualized in the form of beautiful

charts. Textual response questions, on the other hand, should be summarized using keywords (tag

clouds). At an additional cost, clients can choose to utilize an AI (artificial intelligence) based response

analyser that can summarize the long responses into a few sentences. The AI summarizer is provided

via a third party library.

Apart from the actual survey results, most clients also like to review the survey engagement reports.

These reports include visualizations for surveys statistics such as average time spent by participants on

a given question, total time spent on survey, demographics and attributes of participants, and drop-

out reports indicating questions which cause respondents to abandon the survey because the

questions could be too tedious or repetitive.

All the functionality implemented by the information system is accessible through Fynapa’s client and

consumer (panellist) websites. When consumers login, they are presented with a dashboard containing

the overview of their account. This site includes sections on currently available surveys, record of past

surveys and a statement of reward earnings and redemptions. Client site on the other hand has

sections of currently running surveys, archived surveys, payment history and invoicing. Clicking any

completed survey will lead to the results page where the aforementioned reports are presented.