Kissmetrics Blog

A blog about analytics, marketing and testing

Built to optimize growth. Track, analyze and engage to get more customers.

Do More Tweets Equal More Holiday Sales?

With the holiday season in full swing both marketers and economists alike are watching the retail sector closely to understand how the economy is doing and whether they put their marketing dollars to good use. Hundreds of millions of people participate in conversations within social networks each day, and many of these conversations reference companies, domains and products, especially during the holiday season.

Leading indicators of product sales for online merchants have typically involved examining on-domain funnel metrics such as impressions, click through and conversion rates. But, if people were talking about a particular product and you were able to measure the quantity and quality of conversations about that product, then these metrics could be accurate indicators of how well those products will sell. In the social world, the quantitative metric is number of mentions and the qualitative measurement is sentiment. Number of mentions measures how many times a keyword (product) is referenced within a post over time, while sentiment measures how positively or negatively people express themselves about that product.

Focusing on both popular electronics domains and then more specifically at two popular holiday gifts, the Apple iPad and Amazon Kindle Fire, we look to the conversations taking place on Twitter and analyze both number of mentions about each product and sentiment to determine if social conversations are a realistic indicator of Black Friday and Cyber Monday holiday sales.

November At A Glance

Looking at Figure 1 below, it’s clear that references to retail domains in social communication significantly spike in and around Black Friday and Cyber Monday, November 25 and November 28, when sales correspondingly also rose dramatically.

Figure 1 – Social Mentions for Retail Sites In November

Figure 1 – Social Mentions for Retail Sites In November

These mention spikes are commensurate with the retail sales increases realized by online retailers. But, can these measurements be applied to specific products and used as leading indicators of how well products will sell?

Let’s take a look at specific product mentions around Black Friday and Cyber Monday to understand if these measurements can serve as leading indicators of how well products will sell.

iPad vs Kindle Fire

The first analysis uses number of mentions as the base measurement, but expresses this as share of voice, as depicted in Figure 2, for the month of November.

Figure 2 – iPad vs Kindle Fire Share of Voice

Figure 2 – iPad vs Kindle Fire Share of Voice

Share of voice is calculated by summing the total number of mentions of both iPad and the Kindle Fire, and then determining the relative percentage of mentions for the iPad versus the percentage of mentions for the Kindle Fire from the summed total.

Looking at the Kindle Fire, we notice that on 11/13, the week the Kindle Fire became available, there’s a spike in the number of mentions, and then another small spike again around Thanksgiving and Black Friday as the hype began to die down and some negative reviews began to surface.

In the case of the iPad, based upon number of mentions it is still largely the dominant product, particularly among the Twitter community.  And, as the buzz around the launch of the Kindle Fire faded, Apple’s SOV climbed back up, especially with Apple announcing that they would offer free shipping and rumors about Microsoft Office becoming available on iPads

That’s the conversational quantitative perspective… Next, let’s take a look at how people feel about either product as expressed through sentiment. This measurement uses Topsy’s Social Sentiment™, which is an algorithm specifically tuned to measure how positive or negative people express themselves within social media.

Figure 3 – iPad vs Kindle Fire Topsy Social Sentiment ™

iPad vs Kindle Fire Topsy Social Sentiment

Focusing just on the week around Thanksgiving, people express much more positive commentary about the Kindle Fire, with the product’s positive sentiment passing the iPad as consumer conversations moved towards Black Friday. The hype surrounding Kindle Fire’s launch definitely helped propel positive sentiment moving into the start of the holiday shopping season. However, that positive conversation was not sustainable and is attributed to complaints about the product such as “fat-finger” issues making navigation difficult.  And yet, even without a recent product launch, the iPad’s sentiment was also positive, picking up slightly on Cyber Monday as Apple announced that they would offer free shipping for purchases made on

Combining the sentiment analysis with number of mentions, let’s see how these metrics map to one another. It’s evident people are excited about the tablet space this holiday season – but, are all mentions good mentions?  The two charts below show volume and sentiment metrics for both products specifically during the week of Black Friday and Cyber Monday.


Ipad holiday sales 2011

As we saw previously, sentiment for the iPad is positive, staying well above zero with increases on the 25th and 28th.  And yet, while the sentiment still appears to be positive, it is noticeably lower by the end of the Thanksgiving week than at the beginning. However, the amount of conversation around the iPad increases from the beginning of the week to the end, going up by nearly 30,000 mentions. So, does an increase in Twitter chatter really drive positive brand sentiment? In this case it doesn’t appear to.

Kindle Fire:

Kindle Fire Holiday Sales 2011

While Black Friday proved to be a good day for the Kindle Fire in terms of positive sentiment, there were slightly fewer mentions between the 24th and the 26th and yet that’s when positive sentiment was at its peak. Consistent with the trends displayed by the iPad graph, it’s clear that an increase in tweets doesn’t necessarily mean an increase in positive sentiment.

As many predicted, this graph also supports the observations that after Black Friday the excitement around the Kindle Fire began to smolder and by Cyber Monday, sentiment had declined dramatically.  As users got their hands on the device and tried it out, many complained about specific feature dysfunctions and poor design; those complaints were not only voiced on via customer ratings, but also on social media outlets such as Twitter, and as a, result sentiment declined while number of mentions went up again.

Looking back at the sentiment for both the iPad and the Kindle Fire, despite some early frustrations with the Kindle Fire, for the course of Thanksgiving shopping week, sentiment was still significantly higher for Kindle Fire than for the iPad. Even with the dip after Black Friday, sentiment for the Kindle Fire was still on an upward trajectory.  We’d expect that while hype of the Kindle Fire might have slowed, sales should continue to increase, particularly with the OTA update scheduled for the coming weeks.


Quantifying and qualifying social communication about companies and products can provide valuable insights into demand, especially for consumer products where people are expressing opinions within social settings. Measuring the volume (number of mention) and how people feel (sentiment) enable retailers to obtain off-domain indicators for the demand that may exist for the products and provide valuable signal for what products to promote and what priority to display products within their domain.  In this case, just because Kindle Fire spiked in popularity, doesn’t mean Best Buy should move that Apple display to the back of the store just yet. However, the comparably large number of mentions of the iPad shouldn’t mask the Kindle Fire’s strong positive sentiment among its user base and shouldn’t be overlooked as a strong contender for tablet market share either.

What’s needed to “connect the dots” for off-domain measurements and on-domain metrics is to combine the two into consolidated analyses.  One way to execute this consolidated view of measurements would be to utilize KISSmetrics analytics in conjunction with Topsy Labs’ social metrics. This would allow you to combine number of mention, sentiment with page view, click through, conversation rate, average order size and other RPM metrics for each product. The powerful aspect of injecting social metrics into on-domain is that online retailers can utilize off-domain metrics to anticipate product demand, planning promotions and inventory accordingly to stay ahead of and profit from consumer’s social conversations.

About Topsy Labs:

Topsy Labs, Inc. enables businesses around the globe to apply social intelligence to realtime decisioning. Our fast-indexing technology ingests massive amounts of authored content from the world’s largest social networks and our live-ranking software applies influence algorithms to social data, identifying the most important content seconds after it has been posted to the social web. We produce a variety of metrics calculated from census-based social data sets, enabling businesses to interpret social signal in ways that are meaningful to each business segment’s needs, with high degrees of precision and in realtime.

Topsy Labs is located in San Francisco and funded by top-tier Silicon Valley venture firms. For more information, visit or contact us at

  1. Great Article!

1 comment

Please use your real name and a corresponding social media profile when commenting. Otherwise, your comment may be deleted.

← Previous ArticleNext Article →