Microsoft: 2020 Genius Awards Finalist (Marketing Analytics Adoption)
This is the third in a series of 12 blog posts showcasing the achievements of the finalists of the 2020 ANA Genius Awards, presented by Neustar. The 2020 Genius Awards winners will be announced at the 2020 ANA Masters of Marketing Conference, held October 21-23.
CATEGORY: MARKETING ANALYTICS ADOPTION
Sales is obviously an essential measure of business success, but it's not the only one; brand building is also key to long-term gain. At Microsoft, marketers are deeply concerned about driving product perceptions, brand awareness and loyalty, and long-term growth in addition to sales. Here's how the company's data and marketing teams came together to get a holistic understanding of the total impact of advertising ROI, both in the short term and the long term.
Create predictive models that can accurately forecast sales and changes in brand perception
Data engineers at Microsoft had sourced, cleaned and created a pristine data lake. They had built, trained and operationalized a predictive model that could accurately forecast product sales.
At last, they were ready to celebrate their success, but then...
Only after presenting their hard work to their marketing colleagues, did the data science team discover that the insights they had uncovered didn't reveal the full picture the marketing team needed.
The data team's work would help reveal what was driving sales, but these analytics wouldn't shed light on any of the company's other advertising goals, namely driving product perceptions and the overall strength of the Microsoft brand.
"We solved a business problem, not the business," explained Robert Graves, Director, Data Management and Science at Microsoft.
That was a turning point, one where the two teams came together to align on what marketing success looks like at Microsoft—and the best ways to bring together the right data needed to measure the full impact of the company's advertising efforts, not simply the product units sold or revenue generated.
By aligning on the goals, colleagues from both teams were able to build an infrastructure that would truly help the marketing arm of the company, tasked with balancing both short-term and long-term growth, understand how advertising was driving both sales and brand perceptions.
Bring together data insights plus market research to create an upgraded analytics framework
The company had two distinct toolsets for measuring marketing success—the model the data science team had built, as described above, and a separate system developed by the marketing side of the business to track the efficacy of the company's advertising, utilizing custom surveys, digital measurement, and syndicated research. With data going back many years across numerous target markets, the marketing team's system provided access to a wealth of information about customers' perceptions, movement through the purchase funnel, and advertising recognition and exposure.
In combining forces, they were able to create an upgraded data and analytics framework built utilizing the company's Azure Data Lake filled with years of perception trends, mountains of media spend information covering both Microsoft and its competitors, a long history of promotional and discount activity about the brand and the competition, and sales information.
By bringing the best of both of the company's previously siloed systems, they were able to engineer a much-needed holistic view of the company's advertising ROI, one that would help the company forecast both short-term and long-term growth across sales, brand perception, and brand value.
The new framework has enabled Microsoft to predict sales and longer-term factors like brand loyalty
As a result of the collaboration between the marketing and data teams, Microsoft was able to create an integrated sales- and perceptions-based model to predict with reasonable accuracy the amount of sales activity the company could expect to see within a timeframe of 0-3 months of a given advertising campaign. That was a win for measuring short-term impact.
On the long-term side, the data scientists tapped the company's depth of perception data and existing modeling to explore the relationship between perception changes and sales. By digging into these insights, they were able to pinpoint predictable patterns between media spend, uplifts in perception, increased funnel activity, and, ultimately, sales. Utilizing machine learning models, they developed "perception paths" to measure the long-term impact of marketing activities—stretching about 4-9 months into the future.
"This combined short-term and long-term model has allowed us to have greater impact across the business—in media, finance and creative teams," said Graves.
On the media side, this more complete picture of marketing ROI has enabled Microsoft to be more strategic about its marketing investments, guiding the team to support the right media mix and flighting decisions. And while in the past advertising team members hesitated to use the data science team's original short-term modeling to guide decision making around driving long-term perceptions, the company's marketers now have the solid foundation needed to answer questions about the overall customer journey.
For the creative team, there's a greater understanding of how ads showcasing different products impact perceptions across the full portfolio of Microsoft offerings and the brand as a whole.
And as for the finance team, they've leveraged the new model to run simulations to predict spending and create more accurate budgets.
These same simulations have also helped the company's retail business plan for promotional activities and understand the potential impact of the competition.
"As we enter the new fiscal year in a time of tremendous uncertainty, our models provide critical decision-making tools to our executive decision-makers," said Graves.
When looking at short-term implications alone, tactics like promotions and discounts and channels like digital media and lower-in-the-funnel-marketing may seem like the most efficient or effective marketing approaches. For a long time, digital media has provided the easiest pathway to measurability, making it easier for marketers to inherently favor—or more seriously, bias—these channels. To the potential detriment of marketing ROI.
But for Microsoft, these preconceived notions have evolved, thanks to their new analytics framework that illuminates both the short-term and long-term views. Through this advanced modeling, Microsoft more fully understands the aggregate impact of all of its advertising activities and has been able to validate the need for television and perception-based advertising, avenues that are easily passed over when only looking at short-term, performance-based reporting.
Armed with tools that enable the marketing team to measure the impact of all media types and all channels over time and understand how changes in perceptions and brand affect the customer purchase journey, the company is able to make more impactful decisions that boost sales while fostering brand equity.
More about the 2020 Genius Awards Finalists