Google have released SyntaxNet to the public, and as an analyst let me say this: this is the biggest news we’ve heard in a long time. Here’s why.
I spoke at a Link Humans event last week, and it was all about social analytics and “drawing consumer insights from your social analytics”.
Instead of picking just a topic, I thought I’d hold a “social analytics intervention”, talking about a few things I need marketers and social analysts to be aware of in 2016. Just four points, four slides, in 20 or less minutes.
For anyone who couldn’t make it, for anyone who was there but would like a recap, or if you’re just interested to know what I had to get off my chest, here’s what I talked about.
The way you manage your Boolean queries differs from tool to tool. Some tools let you build queries that you can file in folders, like Sprinklr with its topics and topics groups; other tools let you build queries and subqueries, as a way to categorise your search – so for a query looking into Apple products, you may have a subquery for each Apple product (tools like Synthesio support that functionality).
Brandwatch handles queries differently, letting you set various filters (tags and categories) and rules around them to “supercharge” your dashboards.
The problem that plagues all of these tools, however, is that as time passes you end up with so many overlapping queries: you may find two or more queries covering the same topic (thus retrieving the same mentions), or multiple duplicate queries. The likelihood of this happening will only increase as you start sharing the platform with more users.
So, how do you manage queries to make sure that your social listening platform doesn’t get too unwieldy and cluttered with queries? I’ve seen a lot of social analysts use different ways to manage queries, but I’ve settled with one method that works best with me: using a catch-all query.
When doing social listening, how do you track mentions of products with similar names? Or better, how do you do social listening for two products when one’s name is just a variant of the other? Products like the iPhone 6S and iPhone 6S Plus, or the Microsoft Surface and Microsoft Surface Book, or the Lumia 950 and Lumia 950 XL…
There are so many features to look out for when shopping for social listening tools, but one thing that should definitely be on your list of requirements is this: proximity operators. Here’s what they are and why you need them.
One question I recently had from a few work colleagues was about the use of the NOT operator in Brandwatch dashboards, and how that differs from using Brandwatch lists.
If you’re already a Brandwatch user then you may be familiar with their lists: site lists, author lists, location lists. These help you group sites, authors and locations so you can later filter your data based on the lists you’ve curated. (You can find more info about it here.)
However, if you have a dashboard and you exclude authors via an author list or if you exclude them right in the query (using the author: operator), you’ll get the same result: you won’t see mentions from those people anywhere in your dashboard. The same applies when using the site: operator vs. filtering out site lists, and using the 5 location operators (city:, county:, state:, country:, continent:) vs. filtering out location lists.
I’ve seen a few Brandwatch users filtering lists out instead of using the NOT operator to exclude mentions, mainly when it comes to long lists of parameters (authors, locations or sites). For instance, if you have a really long list of authors, you can add them to a list which you can use to quickly filter their mentions out, instead of dumping all of those names in the query, especially if you don’t have enough space left in the query builder.
So, what’s essentially the difference between the two? Are there any advantages in using the custom operators rather than filtering out by lists?
Brandwatch have unveiled a new feature: they’ve introduced a new Boolean operator that lets you retrieve any Tweets containing links to a website, even if the links have been shortened. You can read more about it in their official announcement here.
What makes this operator so special is the second part of its definition: “even if the links have been shortened”. The vast majority of social monitoring tools can only retrieve links as long as they haven’t been shortened by a URL shortener like bit.ly, Buffer’s buff.ly, Hootsuite’s ow.ly, and several more. Brandwatch is now one of the very few tools to retrieve all links, whether they’ve been shortened or not. To illustrate why this makes a huge difference it I’ll use the Brandwatch website as an example.
Quick one: not only am I going to start posting more on brnrd.me, but I’m also going to start posting bits and pieces on Brandwatch – tips and tricks, thoughts and deep-dives on features, as well as new (and perhaps unusual) ways of using the platform (am I the only one using Brackets to write Brandwatch queries?).
You may find this useful for you if you’re already a Brandwatch user (or if you’re thinking of getting Brandwatch). So, to make it easy to find all these Brandwatch posts, I’m going to file them all under this new category: let’s call it the Brandwatch Corner, shall we?
I’ve been reading quite a few articles on “the state of social media” recently. They’re interesting, I can’t help but notice how a lot of these articles seem to ignore analytics, which is a shame considering how great the last 6 months have been for social analytics.
If the first 6 months were busy with changes from social platforms, acquisitions (e.g. Sprinklr, Clarabridge), new platforms coming up (e.g. Meerkat, Periscope), old platforms going through major makeovers (e.g. Pinterest, Google+), new data being made available (e.g. Facebook and Datasift), I can only imagine what’s in store for the next 6 months ahead of us.
I sent this Tweet the other day:
On the lookout for a social analytics tool that doesn't suck…
— Ben Donkor ⚡️ (@FR314) June 9, 2015
To say that it caused quite a stir would be an understatement: it resulted in a few replies, a couple of calls, quite a few DMs and several emails, ranging from pitches to people sharing their own opinions, recommendations, worries and more.
I should really explain what I meant by that Tweet.