June 20, 2018
In The News

Top 10 Predictive Analytics Solution Providers - 2018

The Downward Trend of Trend Analytics

At the start of every season and in every vertical, there exists a small army of prognosticators: industry voices and vendors who claim crystal-ball, trend-forecasting abilities. From media publications (“Trust us, we’re well known, we’ll tell you what the next big fashion is.”) to social listening tools (“Look how many people are joking about this meme right now!”) to traditional research firms (“Sorry we can only project what was in market last year.. we don’t venture to speak about anything new.”), there is no shortage of those willing to add to the pile of guesses on what might happen in the marketplace over the coming months.

Most of these guesses are unreliable, and each of the above examples have at least one severe weakness. The results are either obvious, biased, dominated by noise, backward-looking, incomplete, or have an agenda. And of course, nobody ever goes back to see who was right and who was flat-out wrong, compared to what actually happened in the marketplace. Accountability is something most forecasters avoid.

For casual market observers, this is all fine; results are meant to entertain or inspire, and for them it does not matter if most predictions are wrong. 

But for major brand strategists, who are staking millions of dollars and likely their jobs on trend-related bets, the stakes are higher. For the clients who want to make these bets with as much confidence as possible, Quantifind is trying to be different in order to find the trends that really matter for our clients. Some of the steps we take to differentiate ourselves include:

  • Removing the noise prevalent in all social analytics
  • Building models with machine-learning to classify potential break-out trends
  • Validating any specific trend across multiple independent data sets
  • Describing each trend with rich demographic and psychographic information
  • Discovering and categorizing new trends
  • Ensuring that the trends are statistically relevant to a brand's actual customers

Retail Relevant Examples 

With the holiday season approaching, now is an opportune time to analyze trends as they relate to retail in particular. As a result of analyzing these trends, we attempt to show how Quantifind goes several steps beyond obvious trend results from typical solutions.

Payment Methods:

Everyone knows digital payment systems are on the rise. But with the industry going through growing pains (see cryptocurrencies) and a glut of smaller players battling for position among the major tech companies, it can be hard to assess the landscape. Systematically evaluating trends, we can judge the horserace as seen in the following figure. The stalwart, Paypal, is being eaten at by relative newcomers, Venmo and Cash App. Others including Apple Pay, Google Pay, Facebook Messenger and Zelle, have yet to break out relatively.

Venmo and Cash App have eaten away at Paypal.

All of these trends are non-surprisingly being driven by younger demographic segments; however, the regional differences are striking. Venmo is more common on the coasts (Western and Northeast regions), while Cash App is dominated by consumers in the South. For a retail brand considering partnerships with a digital payment service, knowing these details and scoping the landscape in this way are critical to ensuring synergy in the partnership.

Swimming Against the Downward Physical Retail Trend:

Not surprisingly, our data backs up the general downward trend of many classic retail locations (Sears, malls in general, etc.). Even though these overall trends are expected, by going granular we can find the pockets of activity (or signs of life) within those audiences that are holding on with more strength, such as certain regions or demographics. We can also discover retail chains that are bucking the trend entirely and try to extract lessons to be learned from these exceptions. Uniqlo, for example, continues to enjoy growth, spurred by a popular November sales strategy. Finally, we can learn from what is eating the lunch of most other retail brands. In addition to the obvious Amazon ecosystem, this includes several new subscription service brands (Fashion Nova, Allbirds, among others). By analyzing the drivers and demographics of these invaders, retail brands can at least gain some intelligence by which to stem the tide, and strategically partner or fight back.

These trends are normalized to be shown on the same vertical scale. Certain retail trends are swimming upstream relative to the overall downward trend in, for example, "mall" mentions.

Culture Surfing:

Any general trend analysis over the last few years would reveal several “K-Pop” (Korean Pop Music) related trends. Bands like BlackPink, BTS, NCT, and their cohorts have been rising in international prominence and, while close to the peak of their hype-cycle, have yet to drop off significantly. Apart from the obvious music and merchandising angles, our “related trends” feature reveals more retail opportunities. For example, “lightsticks” is a prominent retail toy trend on the rise that is mirroring the K-Pop trend lock-and-step. See figure below. (This is in contrast to another stick related trend “selfie sticks” which, we can mercifully report, is concluding its death spiral.) While lightsticks may not be the top Christmas break-out seller (they are still dominantly Asian, despite increasing cross-over), they point to an overall polarization and tribalism trend across music, sports, and, obviously, politics. “Fan Armies” attached to music artists and “Stans” in athlete-following culture, indicate the increasing degree of tribalism that can be monetized, but only if the brands in their orbit recognize the relevant opportunities.

Both curves are normalized to be shown at the same scale, cleaned volume from social media. Note how Lightsticks volume precedes certain events.


Summary

The beauty of the unstructured data sources that we use for trend analysis is that they cover most verticals, from retail to restaurants to financial services, using a very broad base of consumers. This lets us observe cross-vertical and cross-trend relationships and build a graph that gives a holistic, forest-for-the-trees, picture of the entire market. Unlike other datasets which break trends into isolated silos, these interactions represent where many of the best strategic opportunities will be for brands in coming years.

The examples above only touch the surface of Quantifind’s collective trend knowledge-base. With our categorized database of 40,000 (and growing) trend candidates, we can create industry reports for any vertical, with granular insights on their compositions and relationships. 

Given the meta-trends of 1) an increasing amount of noise in social measurement channels, 2) the accelerating speed of the marketplace, and 3) the growing complexity of interaction between trends, brands must use the methods discussed here to even hope to stay strategically relevant.

To learn more about how Quantifind trends can help your business, please contact us at insights@quantifind.com.