Heineken sponsored a CrowdAnalytix contest
to analyse their marketing efforts during the US Open.
The US Open was held between Aug 26 and Sep 9, 2013. This is a visualisation
of the first dataset that has daily information for 29 days (from Aug 19 to
Sep 15) on various social media metrics.
Below is a hierarchical clusterplot showing the Twitter metrics alone. Each
cell shows the pairwise correlation between two metrics. Each cell at the
intersection of 2 metrics has a value, e.g. 86, that indicates an 86%
correlation between the two metrics. These are coloured green if they show a
high correlation, or red if they show a low correlation.
Based on this, we can identify whether any metric strongly affects any other
metric.
“Correlation is not causation but it sure is a hint.”
Insights
A. For Heineken and Budweiser, the number of official tweets, retweets and links shared are highly correlated — understandably.
B. ... but the number of replies is unrelated to these.
C. But while Heineken got a lots of retweets (Retweets received = 86%) when the official responses increase, Budweiser did not (37%).
D. Official replies are completely uncorrelated with any other metric, particularly the number of retweets received. So replying more does not appear to be a good *viral* marketing tactic. (It may make sense for other reasons.)
E. Budweiser's tweets had more impact on its hashtagged tweets than Heineken's. What did Budweiser tweet about that was more effective? Might be worth investigating.
F. For both parties, the number of influential tweeters is perfectly correlated with #tag tweets (97 - 98%). But then, maybe that's the definition of an influential tweeter?
Full metric list
The full list of metrics also includes data on the number of Facebook posts,
videos, etc. along with Google trends data. The structure below shows the
correllations across all of these, including the Twitter metrics.
Insights
G. Retweets received for Budweiser seems to rise with their FB videos posted.
H. These retweets also seem to be strongly correlated with Google Trends for Budweiser, but not for Heineken
I. While Budweiser #tags and retweets are influencing their Google trends worldwide, whereas Heineken's are only influencing the US.
Timeline & trends:
Another area of interest is how Heineken's performance differed from
Budweiser's on a day-on-day basis. For example, were there specific days
when one firm's tweets shot up well above the other?
The visualisations below show the metrics for both Budweiser and Heineken.
If Heineken does better, the difference is shown in green.
If Budweiser does better, the difference is shown in red.
Insights
J. On Sat Aug 24th, two days before the US Open, Heineken's Facebook posts saw a significant spike in shares (1,800+) and likes (39,000+).
K. On Wed Sep 4th, Budweiser's Facebook page likes shot up dramatically. This does not appear to be driven by any specific photos of videos posted on their page, however. Two days later, the number of FB page users also shoots ip dramatically -- but only for a week. After that, it dies down.
L. On Sun Sep 1st, Budweiser tweeted several times, mostly with links.
These visualisations were created using the Gramener visualisation server.
See the [interactive HTML5 version](https://gramener.com/usopen/) with filters on the website.