Microsoft

Six years ago, around the time when Microsoft acquired the mobile phone business of Nokia, the team founded in Microsoft got a mission to take ownership on the market research data of the whole company. Having already worked together at Nokia, the posse rolled up their sleeves and built top-notch data analyzing processes based on their previous learnings, to help the decision-making of executive managers at Microsoft.

The core task for the team is to help the decision-makers at Microsoft, who often are marketing and product directors, to get up-to-date information on the wishes, behaviour, and opinions of consumers to support sales and target setting.

“The survey data coming from market research vendors is very raw, so it needs to be “cooked”. The aim of the team is to make the raw, scattered, and sometimes even partial data into as understandable and current as possible for business decisions”, says Matti Uronen, Senior Data & Applied Scientist of the team.

The team has thus developed effective tools and processes to process and confirm the validity of the research data coming from the vendors, to forward it to cubes, and finally get it into digestible reports to various dashboards and analysis services, such as Power BI, and to databases.

Outsourced market research data can be enriched with the internal data owned by Microsoft, and this makes the scale of analyzed data advanced compared to other companies.

The entity of tools and processes developed by the team is extremely efficient despite the huge amounts of data, as the processing and curating of the raw data file coming from a market research vendor, and then making it available for the Microsoft personnel to read in the forms of metrics and KPIs, is mostly automated and can take just a couple of hours.

“When the data analysis is ready, we will send a notification to our users on the new data being available. After this, our decision-makers examine the data directly from the databases or e.g. from Power BI. Compared to static PowerPoint slides the decision-makers can directly create comparisons, edit values, and explore with the analytics tools how different choices effect on values”, Saarikko describes.

Additionally, the team has developed new, innovative ways to combine different types of data and to create new kinds of metrics by taking advance of Microsoft’s own products, especially Azure.

“We have yearly projects where our internal customer needs new kind of metrics or there appears a need for us to produce text analysis in an easier way. Our team is flexible, so if someone wants new KPIs or data in their use, we think how to do it and how we can take all out of our tools”, Saarikko says.

“Microsoft has collected understanding, according to the terms of use, of the profiles of its customers, so innovation can thus mean enabling better samples. It is very different for results to interview customers of whose behaviour and opinions there already exists understanding, compared to researching the views of completely random people. For example, we can do a targeted surveys and give decisions-makers information on whether marketing or some product’s features need some development. We also do plenty of prototypes and our projects don’t last for months or years – instead, we react quickly”, Uronen continues.