How to manage complex multi-country research projects
Running multi-country studies is a challenge and we know that from experience! What problems have you noticed in the manner these studies are currently handled? Have you wondered if there is a way to increase their efficiency so that you can use them to build even better strategies? GemSeek runs more than a million interviews in 50 markets every year and here’s some advice from us.
Ensure the consistency of the data generated from the research by streamlining project management and setting clear rules for survey execution. A good approach is to create a survey manual. It will describe the steps to be followed and allow you to have the same process in place across different countries. It should outline team roles, as well as all processes – from research design and sample definition, to data processing and reporting.”
Next, you should set KPIs to establish a baseline, estimate the performance of each country, and track the progress, if it is a multi-wave tracking study. Don’t forget to use metrics that are meaningful and easy to explain across different stakeholders. Purchase Funnel, Net Promoter Score or Consumer Decision Journey framework are well understood and proven to have correlation with market performance.
Once you are done with setting measurable criteria you should carefully define the quotas in each country. We recommend self-completed online interviews for the multi-country projects. Keep in mind that in countries where internet penetration is still below or around 50%, achieving a nationally representative sample might be a challenge.
Being realistic makes all decisions a lot easier. Your company’s internal stakeholders can be disappointed when they have unrealistic expectations. That’s why you should engage them from the beginning of the planning process, where you can agree upon realistic goals and targets. It is important that all of them have the same understanding of the main purpose of the research project so that they can be aligned towards this one common goal.
We all have deadlines and it is highly possible to find yourself in a hurry to kick-off the research as soon as possible. However, to ensure a smooth start and avoid mistakes, we recommend soft launching the project in one pilot country where you can test the methodology and the process. After that you will surely be more confident to deploy the survey in all other locations. Oh, and if you happen to encounter any problems, you will have the chance to fix them before the full release of the survey!
Small details can make a big difference
Master the research materials
Master versions of all survey assets (questionnaire, brand list, programming script etc.) are a prerequisite for a successful multi-country project. Create one master questionnaire but leave room for location specific questions to address the key stakeholders’ needs. Ideally, these add-on questions come at the end of the questionnaire and are no longer than 5 minutes, guarantying that the rest of the survey won’t be influenced and the output will be comparable across countries. To achieve even greater comparability, we advise you to double-check the meaning of the questions after their translation. Here, the best thing you can do is to have them checked by native speakers. A master version of the brand list and the link programming script will also serve you well.
Always start with the master version of the research document, and once you’ve agreed on it, continue with the country specific versions.
Fill the missing values
Based on GemSeek’s experience, questionnaires longer than 20 minutes lead to respondent fatigue, thus diminishing the quality of answers. Image statement grids are the types of questions that add the most to the overall questionnaire length. You can optimize those questions and shorten survey length by using data imputation methods to fill in the missing data points. For example, we at GemSeek, employ methods from the data science domain in order to best predict the value for a certain respondent and question. Using mathematical and statistical algorithms such as Conditional Random Forest and AMELIA, responses are accurately ascribed to some respondents based on the overall data of the studied population and without the need for the respondents to go through the entire survey.
Be careful with data aggregations
If you run a multi-country survey you would most probably need an aggregated global data of all the countries involved. In order to reflect the size of each country, you should apply a weighting factor. The weighting factor could be a general one (such as GDP or population) or a company specific KPI (revenue, sales etc.). Choose carefully, as each weighting procedure has its specific impact on the global results. Weighting by company revenue, for example, prioritizes your brand results, giving higher weight to countries where you have higher sales value.
Increase the efficiency of coding and charting by automating the process. This will also save you time and money. There are plenty of agencies offering ready solutions which will help you optimize the process while bringing even higher quality.
Be cautious when analyzing the results
Metrics won’t tell you much, if you fail to acknowledge cultural differences. In India, for example, respondents tend to give more positive answers than in Japan. This doesn’t mean that the given brand or product performs worse in Japan than in India, but just that the analysis requires understanding of local specifics.
Engage the stakeholders:
Organizing teleconferences where you can present and explain the key findings is a must but you shouldn’t stop there. Engage your stakeholders further to ensure that the findings are properly interpreted. Create dashboards with all the data you need at your fingertips. If you manage a tracker survey, you might consider sending out regular newsletters to provide updates and useful insights.
Multi-country research may seem overly complex, but the benefits from it outweigh by far any disadvantages – it brings more value for money and provides a solid base for benchmarking. To turn this seemingly complex process into a smooth and productive experience, just make sure to plan SMART, pay attention to small details and speed up the analysis without compromising the quality.
Do you have questions? Get in touch and one of our experts will contact you with customized advice!