Reaching the Respondents

The target respondents of the survey were selected in parallel with the aim of the research. Employees with significant job titles who work in Turkish firms, which have international activities, are the main targets to in order to get accurate responses from the people who can be related with the research with their area of profession. The first source for reaching out respondents is people that are reachable without any mediator person or computer program. These respondents are family, friends and people that the survey can be given by first hand. The second source of reaching out respondents is the second hand network. These respondents are family and friends of the first hand network. First and second hand network are good starting points to receive initial results. It gives you an idea how the responses will be shaped like. However the number of responses wasn’t enough that is received from first and second hand network in order to finalize the survey results. Therefore a third source was needed in order to finalize the results and moving on to the analysis part. For the third source, after many research and discussion, the internet/app service LinkedIn was chosen. As it is known, it is an employment-oriented service and there are many employees that are registered from many different countries and companies with many different job titles. In order to target respondents there are many filter options on LinkedIn that the respondent pool can be narrowed down. For the filter options; company and job title had to be determined. Before proceeding on LinkedIn, company research was made. The most active Turkish companies in foreign markets were determined.

After generating the company list the job titles was determined. It had to be relevant job titles and seniority of employees in order to get accurate responses.

On LinkedIn’s filter options, Boolean search is possible. In order to put the job titles and seniorities as Boolean formula, key words was determined.

33

• Some of the key words for job titles: Marketing, Sales, Project, Strategic, Operations, Logistics, Brand, Product, Export, Supply Chain, Global, Regional, International, Culture, Country

• Some of the key words for seniorities: CEO, CFO, CMO, Head, Executive, Manager

After the selection of job titles and the companies, by the help of the filter option, the employees, who are the possible respondents, were monitored. The number of monitored employees is different for each company due to the size of the companies and number of registered people on company’s LinkedIn page. Once the monitoring was done for each company, the employees were extracted from the page and put it on an Excel sheet in order to be sorted. The sorting process was mostly about removing the employees that are irrelevant in terms of job title and the company. Even though the filter is used, sometimes there can be a leakage and it had to be removed. Each possible respondent was put on Excel sheet with name, company name and job title. The next step after putting the possible respondents on Excel sheet is to find their email address. On LinkedIn, the email addresses don’t exist. Therefore another method was generated. From the companies website’s, the domain of the email address that the company is using was found.

The domains were found mostly in the “contact us” page of the websites. However, not necessarily each website has the email domain of the company so different kind of sources were used. Once the domains were found the next step was the email format. Again the websites of the companies were the main source along with other Internet sources. Each company has different type of email formats like the email domains. The most email formats are listed below.

• (Name)(Surname)@domain.com

• Name.Surname@domain.com

• Name_Surname@domain.com

• Surname@domain.com

34

• First Letter of Name.Surname@domain.com

• (First Letter of Name)(Surname)@domain.com

In order to accelerate the process, Excel formulas were used to generate the different types of email formats, like the ones listed above, when the name and the surname of the employee is written on the excel sheet. This method was very crucial to reduce the time of the process and once the email format was determined, it was taken from the formulas and combined with the email domain.

At the end every employee’s email address was obtained. Of course there were many obstacles during the process of generating the emails. For example people with middle names was a big obstacle because it gives a possible to email addresses. One of the names or both of the names can be used for email address.

Another obstacle was the women employee with multiple last names due to marriage. In this case, their last name was a problem. It was difficult to determine which last they are using for their email address. The cases like these multiple last names and names increase the chance of variety of the email addresses and it becomes harder to guess the real email address that the employee is actively using.

For this cases email verification websites were used. These verification websites shows the users if the email is valid or not when the full email address is typed in to the verification button. This websites were very crucial tools in order not to skip any email address. After the verifications were done for each respondent candidate the next and final step was to send out the emails. For sending out the emails, VSE’s email address with “vse.cz” domain was used. Before sending out the emails, the next step was designed in advance in the sense of awareness that the respond rate would be low. In survey applications, it is well known that it is not easy to get high response rates. The main reason for that is the survey can take more than 5 minutes and it is very likely that people who receive the survey might not complete or even ignore it. In these conditions, it was necessary to remove every kind of possible barrier that might prevent the survey to reach to the possible respondents. In order to accomplish this, some small touches were made. For example, the emails weren’t sent out at once but it was sent out company by company. The main reason for that is the email server isn’t able to send out more

35 than 500 emails at once. The second reason is that if possible respondents from different companies are put in the receiver segment all together, that might cause the receiver to perceive the email as a spam and that causes a zero response rate.

In order to prevent this problem, the emails had to sorted differently according to a criterion. Sorting the email addresses according to the companies was the most logical way to differentiate them. Another method that was used to enable a high rate of response was to arrange the time of sending the emails. The emails that were sent out during the working days were preferably sent in the mornings or after lunch break. The aim of this action was to increase the chance of receiver to see the email and in parallel to that the survey link attached in the email before any other email messages were sent and blocks the one that has the survey. The same method was used for the weekend with the same logic and the emails were sent out Sunday evenings where the receiver can see the email on Monday morning. By thinking of these small touches, it was aimed to keep the response rate as maximum as it can be, in an environment where it is difficult receive full responses from the people whom the survey is sent to.