Podcast: Navigating large data shifts with Kim Short of Ipsos

On our podcast, Kim Short, Senior Vice President, Ipsos North America joined us to discuss what to do when a longitudinal study is hit with a monumental data shift. Having worked in the industry since 1996, Kim has worked in various capacities, including President of the British Columbia Chapter of the Marketing Research and Intelligence Association, and has been with Ipsos since 2003. She has a passion for understanding what makes brands successful and how to drive their success to new heights with the support of communications that entertain, inform, and persuade.  

During the show, we talked about how data shifts are something that we’ll never escape, and that we’ll never be able to easily predict when or how they will happen. However, that shouldn’t stop us from planning for them and knowing what to do in the likely event of them happening in the future. Having been with Ipsos for nearly 20 years, Kim has seen her fair share of data shifts and their impact on longitudinal studies. 

First, we dove into a high-level definition of a data shift, which occurs when we observe a sudden change in results that falls well outside what has historically been observed for a given measure. As you can imagine, this has happened quite a bit over the past couple of years. Kim says that we generally see two types of data shifts: 

• Internal data shifts occur when there’s been a change in the market research process itself.  This could include things like migrating to a new vendor when a client is making a change to scale a study or because they want to take advantage of increased expertise or a new method they can’t get from their existing vendor. It can also include changing the brands that are tracked or changing things like survey question wording, context or how a question is displayed to respondents. Also, the priorities of the company or brand itself may change. Or the panel blend can shift due to changes how panels are being managed or a move to a new vendor. 
• External data shifts occur from things outside of our control as an organization. It includes more broad changes that result in an overall category disruption. Kim brings up an older example: the introduction of the iPhone into the cellular phone category, which moved everyone away from typing with their phones to swiping. An innovation like this, either on the product or marketing/advertising front,  can transform an entire category, for good. If you don't take this broader view to try to understand these dynamics, your data may stop making sense. Other external shifts, such as an increase in fraud or a worldwide pandemic, can also affect data.

Understanding the source of the shift, no matter which type, is what we need to understand in the context of longitudinal studies. 

Kim also talks about what to look for so that you know there has been a shift. There is a normal up and down of data over time, but we need to look out for very sudden changes or a significant change that builds over a relatively short period of time. This shift could happen over a few quarters or a year depending on how established the category is, the competitive context and the nature of what is causing the change. Internal changes can usually point to a more sudden shift and external factors are often more gradual. It's not always clear, so Kim goes into some examples of good indicators to look for so you can stay aware of any data shifts that will affect your longitudinal study. 

She says if you do see a data shift, first of all - don't panic. Explore the cause and work methodically through where it came from and how to address it. Kim says to check and recheck everything and consult with your team, especially others with more experience. When you keep digging you can find the root of the shift.  She also touches on how, and when, to communicate these changes to stakeholders both internally and externally. 

Finally, we cover the role of technology in helping an insights team move forward with their longitudinal study. Listen to the podcast for some real-world examples and insights into how to best manage your longitudinal studies in the face of data shifts.