Blog

I’ve recently been using analytics to stack two of the market’s top smart phones up against each other – the Apple iPhone 6 and the Samsung Galaxy S6.  My previous blogs dig into the review scores, review topics married with scores, review sentiment by topic, and topic detail.

I’ve already peeled back the layers on the scores and the topics to get to a granular level of detail, and it has appeared throughout my analysis that the iPhone is beating the Galaxy in nearly every category that I’ve compared them on.  For this blog I wanted to dig deeper into the sentiment of the iPhone reviews – I’ve already seen that the sentiment of the consumer’s reviews is positive overall, but are they just merely satisfied, or are they so excited that they can hardly contain themselves in their reviews?

Data

For this analysis I had to look at the data from a different perspective – digging into the intensity of the sentiment.  I decided to choose the topic of “features” to explore further, as it had an overwhelming positive response rate of 83%, and the features of a smart phone mean a great deal to one’s overall experience.  We already know that 83% of the comments regarding the features topic are positive – but there are still some unanswered questions – how positive is it, and is it trending up or down?  Typically, doing an analysis like this could take weeks to put together, but using the power of PolyVista’s automated engine, I can produce reports on this data in a very short amount of time.

Analysis Results

More information about the “Sentiment Intensity Chart” can be found on the ChartExpo™ website.

Interpreting the results

The sentiment intensity chart is pretty simple to interpret, and tells us a pretty good story as well.  To interpret the chart, you can simply look at the trend line and see which way the sentiment is moving.  Each vertical line on the chart is for a time period, in this case I viewed it week-to-week, so we can see how the sentiment is changing by week.  For each week listed on the chart, we can see how positive or negative the responses were for that period, identified by a point on the chart, in either a red or green dot.  The top half of the chart represents positive sentiment (green) and is shown on a scale of 1 to 5, while the negative sentiment is shown on the bottom half of the chart and is shown in red, also on a scale of 1 to 5.  The more intense the sentiment, either positive or negative, the higher number score that period will receive, as well as a darker shade of red or green color also indicates a higher score for that sentiment.

The last three points on the chart are indicated by P1, P2, and P3 – what this is doing is adding a trend line through the data points and predicting out what the next three time periods (weeks in this case) will look like.  We can see in this chart that sentiment is trending downwards, and is nearing the point of changing from positive sentiment to neutral, and possibly even creeping down to negative.

What does this tell us?  This tells us that for some reason, when the iPhone just came out that people seemed pretty darn happy when they talked about the phone’s features, but as time progressed, the positive sentiment is waning.  The next question you probably have, is why is the sentiment trending downward?  That’s a great question, and the same one that I had.  From this visualization we can see that the sentiment is trending downward – but get no insight into why it’s headed in that direction.

From this point I decided to look at the same data, but from a different angle.  I want to look at how the subtopics tied to sentiment of the features topic are changing over time, and from here I can discern why exactly the sentiment is trending downward.  Again, this is something that is very cumbersome and time-consuming to produce, but given the PolyVista tools within my reach, this chart was produced with minimal effort.

More information about the “Comparison Chart” can be found on the ChartExpo™ website.

This is another chart that is very simple to read, yet tells a pretty powerful story.  The way to read this is that the duration of reviews was divided into five time periods (D1, D2, D3, etc) as noted above the columns.  We can also see how many reviews there were for each time period noted next to the period, for example D1 had 342 reviewers talking about features, while in period 2 that number dropped to 220.  Working down each column it’s easy to see some of the sub-topics within the features topic, again with counts and percentages related to that period.
Within each subtopic, we can also see a red or green percentage, and this is the percent of the comments from that subtopic that fell into the positive (green) or negative (red) sentiment bucket.  Only the greater number is shown, so for example in D1 for finger scanner we see 68% of the comments are positive, and can assume that the remaining 32% are negative (100%-68%=32%).  We can also see how the topics flow from period-to-period as they are connected with bars of the same color – this is valuable to see how subtopics are trending over the full period – are they going up/down or staying at the same place.

What I see here are a couple of things – the first that sticks out is that the comment counts are diminishing over time.  This is to be expected; as with any product launch it’s the hottest when it first comes out, but over time the chatter tends to slim down.  This falls right in line with a fact I noted earlier relating to iPhone loyalty, as users are jumping to get the newest model right when it comes out.

A couple of areas that I can easily spot as trending downwards are for the finger scanner and iCloud.  For finger scanner, this subtopic averaged 28% negative over the first three time periods, and jumped up to 33% over the final two periods.  For iCloud, over the first three periods averages 32% negative, while over the next two periods the average jumps up to 40%.  There is also the alarm subtopic that pops up higher in period three with 100% negative – also a contributor to the diminishing overall topic sentiment.

From here we have a baseline hypothesis that is supported by the data that over time since the release date of the iPhone 6, the sentiment of the above noted topics has decreased.  I can also dig into the details for these subtopics as I did in my previous blog, but I’ll spare you the gritty details since I’ve already shown you this before.

The Challenge and Opportunity

This is a very interesting challenge, because once you have purchased and own an iPhone, it’s hard for Apple to update the technology and add new features to your device, outside of the annual operating system software update that Apple pushes out (it’s only optional to install the upgrade).  We all know that in today’s world, technology can change at the drop of a hat – what’s innovative and relevant today could be dated and obsolete next week.  When I made my decision on which type of phone to purchase, the features offered was a major selling point, but then like many of you, I tend to get disappointed in a couple months when the technology has evolved and a friend purchases the “hot” new model of phone that just came out and it offers new features that go above and beyond what my device has.

This is an interesting challenge for Apple to conquer – can they figure out how to keep the latest technology in the hands of their users?  Another point to consider is with their known loyalty, do they even care?  It’s known that their loyal users will rush out to buy the newest models right when they come out, so is this a challenge they would even care about?  The answer is yes – keeping your customers satisfied to the highest levels will do nothing but propel the business forward even further.  Should users of other devices see how Apple can keep users updated with the latest technology, then they can capture even more market share, and grow their lead as the top dog in the smart phone category.

Conclusion

I wanted to investigate the intensity of the sentiment for the features topic based on what consumers were saying in iPhone 6 reviews, and noticed that the sentiment is trending downwards quite a bit over time.  To investigate further I created a comparison chart to see exactly how the sub-topics and sentiment has changed over time, and noted a couple of areas where consumers were less satisfied as time progressed.  This creates a pretty unique opportunity for Apple, who has been masters of innovation – can they find a way to keep the latest and greatest features on their user’s devices, and capture even more market share to grow their lead?