Don't Confuse Popularity With Virality

Though suffice to say if it's viral, it's likely popular.

Posted 1 year ago by Tom Maiaroto.

I continually research the area of viral content across the internet. It's something that's pretty fascinating to me (obviously) and at the same time fairly confusing and limited in terms of mass understanding. That is to say, not enough people research it and there is not enough information out there about it.

We have a lot of people and services trying to make things viral...But the jaw dropping part is that they don't actually have a way to measure or quantify their success. Other than perhaps a loose number of increased page views or mentions on Twitter or something. It's very vague and I think many people like it that way because it's easier to for them to justify various things or "prove" their value.

Virality is brutal. It is very unforgiving and the truth of the matter is most things aren't really that viral and/or viral for that long. So people shouldn't expect every single piece of content to be a hit.

Back to my point of the day here. In my continual research I stumbled upon a blog post from Imgur. It's labeled "virality" but in reality it goes on to talk about "popularity" and it is about popularity. The writer in this case is confusing and blurring the lines between the two, but they are completely different things. Further, the way Imgur measures what is popular (and how they rank content, much like Reddit, based partially on votes) is not a measure of virality.

I do want to note that it's an interesting algorithm for ranking and most definitely serves their needs well (I agree with the approach they've taken). No dispute there. I just want to make sure people understand the difference.

This is very important: Virality is the spread of information across the web. The act of upvoting or downvoting content does not spread information. 

Therefore such values have no place in an algorithm designed to measure virality. Not only that, but you're also factoring in metrics from a single source. If they care about "votes" then why aren't they pulling any that might exist from Reddit in? Because it's not their site and they don't care. Which is why their algorithm here is designed specifically for them. It is not a general measurement. Nor is a measurement that you can employ to determine the virality of any other piece of content (not on Imgur). There is, for example, no up/down vote for Twitter and likely not for your average blog.

You also want to be careful not to take into consideration comments. Comments also are not a factor in spreading information. It is, again, a sign of popuarlity. One could fairly use the number of comments in the very same manner as up/down votes. In fact, for bonus points, you can analyze sentiment and use negative comments as down votes.

Hopefully the difference is pretty clear at this point. So, why did the writer use the language "virality" then? Likely because it's a buzz word that would garner some attention. Or perhaps a legitimate misunderstanding or mistake. 

However, the two phenomena often go hand in hand. If people share a video on YouTube or a Reddit post with great frequency, then the content is likely to have a higher number of votes or star ratings. Of course this could be a negative value or a low star rating. Remember just because people hate something doesn't mean it's not popular. It just means it's unfavorable.

Likewise, if something is getting a lot of votes on Reddit, then there is a good chance it's viral. There is a great chance many people are sharing it with their friends. Yes, through Twitter, Facebook, etc. but also through e-mail (which we unfortunately can't really track when it's not our content).

So they go hand in hand, but there's a very distinct difference. Virality Score's algorithm has been specifically designed to measure how viral content is across the entire internet. It's an adaptive algorithm as well. Not only does this increase accuracy, but there's really no other way to be frank. Any algorithm designed to measure something that's in constant change must be designed to be adaptive to the changes in the internet. It must also take into consideration more than just a single site or service.

Filed under
comments powered by Disqus

Search for Posts

Popular Labels

social media virality score general internet web development reviews and opinions php web tools hosting lithium web design

Recent Posts