Interview: How to Use AI for Social Listening With Ben Donkor
By Phillip AgnewNov 14th
Published July 11th 2016
I was pretty excited when Brandwatch released Signals last year, so excited that I started using it for pretty much all of my Queries when it was first released.
I started getting mixed signals from it: some useful, others not so useful, others repetitive. After a while I admit I began to fall out of love with it, starting to doubt if it was useful in the first place. Turns out I was using it completely wrong.
After a bit of work and reorganizing, I finally found the perfect arrangement for Signals.
Here’s how I made this Brandwatch feature my best analytics companion.
The Signals feature doesn’t come with a manual, mainly because it’s so straightforward and intuitive that you don’t need instructions to get started.
However, I just assumed that it could work when attached to all my Queries – but I was wrong. By spreading my signals so wide, a lot of the Signals emails I received weren’t so useful.
So, if I were to give advice to people using Signals for the first time, or trying to make the most out of Signals, these are the three best things you can do.
I realized where I was going wrong, took a look at those three points, fixed the issue, and had Signals work perfectly for me ever since.
I’m going to explain how I’ve built a clear structure around Signals to make them really useful, and show you how you can build your own structure.
First things first – create a catch-all Query.
A “catch-all” Query is your master Query, one that contains everything about your brand. This includes,
Your catch-all is most certainly going to be your biggest Query. Will this replace your existing Queries? Possibly, and the more it replaces, the better.
Why? A lot probably overlap, and unless you have unlimited number of mentions, you may want to streamline how your mentions get pulled into your Queries, making sure that you’re building and managing them efficiently.
Before deleting anything, make a note of your existing Queries, and after you’ve built your catch-all check those that you’re safe to delete. Most of these will become rules that you can attach to categories or tags.
If you’re ready to get started with your catch-all or you’d like to read more on how to maintain it moving forward and how to make it work for you, I’ve got all the info you need in this blog post.
Now that you have your catch-all in place, let’s supercharge Signals.
Signals work best when you attach them to something structured. While you can use Signals for whatever you please, here’s a structured way to apply them.
The catch-all Query you’ve just created includes at least six of these aspects:
Brand names: brand name, and any other names people may use to refer to your brand and trading name.
Brand products: any products owned by you, both current and discontinued.
Brand services: any services you provide, including customer services, consulting etc.
Brand properties: any mascots, buildings and other properties owned by your brand.
One brand property for British Telecom would be their iconic BT Tower, which has its own Twitter account (@bttowerlondon). Online properties are included too, so this includes any website(s) owned by your brand.
Brand social accounts: username of all your social accounts, regardless of social network.
Brand people: names of people in your C-suite, as well as any influential people working for the company or related to it.
For instance, the catch-all Query I have for Microsoft includes Satya Nadella (CEO), as well as Bill Gates and Steve Ballmer only when mentioned in the context of Microsoft – (“Satya Nadella” OR satyanadella) OR ((“Bill Gates” OR billgates OR “Steve Ballmer” OR clippersteveb OR “steven_ballmer”) NEAR/10 Microsoft*).
Create a category for each of these items, and attach a Signal to each category.
Once you’ve done that, you’ll have the following; the most important Signals you’ll ever receive.
These signals include notices of a sudden increase in your brand being mentioned (sometimes a great positive PR opportunity, sometimes a PR nightmare about to happen).
It’ll also include mentions of your brand by influential people, like journalists covering you, a celebrity mentioning you, or someone with millions of followers suddenly slating your brand to their big social audience.
If you don’t already have one, I suggest you create a category called Products and have a subcategory for each product you cover.
Include current and past products (or feel free to keep them in separate categories).
Make sure you also include the social handles of these products, and any monikers that people use when referring to your products.
Now that you have this, create a Signal for this category, and you’ll start receiving noteworthy alerts about your products.
This structure is similar to your Brand Signals, except you’ll have a much better view when you keep brand Signals separate from product Signals.
These Signals will alert you of critical aspects of any service you provide, and it isn’t limited to mentions of your customer service.
Think of any service that you provide to your customers: if you’re a telecom operator then make sure you include mobile signal, 3G vs 4G, mobile data, roaming, tethering etc.
If you’re a catering company, don’t forget about delivery, collection, ordering, booking etc.
Some of these may be so obvious to you that you may not even think about them, but your customers may be having positive or negative experiences with them.
Each and every one of these services affect customers, old and new, and they may also be aspects that prospective customers may be wondering about.
Think, for instance, about people mentioning your brand vs. a competitor (which may also be their current provider) in relation to one particular service.
While it’s important to keep an eye on these mentions and act upon them when needed, it’s extremely important to have a clever alerting system that gives you a heads up every time your services are mentioned in an unusual way, or by more people than normal.
That’s where Signals comes in.
When I say “brand properties”, I’m including stores too, both physical and your website and any digital estates you have.
Here’s a quick tip – make a “Failure and Downtime” tag.
This lets you filter down to mentions of downtime and service failure.
The rule that feeds into this tag needs to include generic keywords that people use when discussing downtime (e.g. down, downtime, crash, “stop* NEAR/0f work*”), service degradation (e.g. unresponsive, slow, “keeps loading”, faulty), as well as any critical state (e.g. hacked), any errors, be they generic (e.g. 404) or custom to you.
If you have any custom text that you display when your system/service goes down, make sure to include that.
Last but not least, include any modifications of your brand/product names that people adapt during downtime.
The most common modification is appending the word “down” or “fail” to your brand name or product name (e.g. #netflixdown, #BTfail).
This tag will always be under construction: people will always come with creative and colorful ways to describe downtime of your service – it may be a hashtag or a new trending expression; make sure you add all of these to your tag, as people may reuse them in the future.
If people start mentioning your social accounts in droves, you’ll want to know about it.
If journalists start covering a story that involves your social accounts in some capacity, you’ll want to know about it.
If it’s related to your social handles, then your community manager or social media manager may already know about it.
However, that may not always be the case – think of out of hours situations. Setting up a Signal for your social accounts helps you stay on the pulse without constantly checking your social notifications.
In fact, from personal experience I know that Signals works a lot better than the daily/weekly summaries that Twitter offers and the notification emails from Facebook.
While these social networks offer you mail about what you may have missed, Signals offers you timely alerts on what you definitely can’t miss.
Your brand isn’t only made up of your products and whatever else you offer.
The people in your C-suite are included, and so are any influential or active employees working for the company.
Oftentimes people who used to be part of your senior leadership may still be mentioned in relation to your brand, even if they don’t work there any more – think of all the times Bill Gates gets mentioned on Twitter about Microsoft, even though he hasn’t been the CEO of Microsoft since 2000.
Most of the alerts I receive for this are actually mention from the press, like when our CEO makes a new announcement. This is great for your PR team, and I’d suggest including them in this particular category of Signals (set them to “high priority Signals”).
That way if something’s being said about someone senior or influential in your company that needs their immediate attention, they’ll be able to draft a PR statement to go out in a timely manner, rather than waiting to hear the news from a slow chain of emails.
As senior people and executives leave and join your board, make sure you keep your category up to date.
There’s a seventh category of Signals I recommend having: event signals.
If you have an event you’d like to track, or any company event with a hashtag and you’d like to make sure you’ve covered all your stops, create an event signal.
This type of Signals goes beyond actual event, and it covers anything that might have happened around the brand: from a new launch to a redesigned logo, from a new CEO joining the board to an outage.
Now, how do you manage them?
In your email client of choice, create a folder called “Signals”, and create subfolders based on the main categories of those you’ve created.
This makes it more manageable, and if you’re in Outlook more than you’re in Brandwatch, you’ll find this to be a great way to still be on pulse should anything happen that requires your attention.
If you use Gmail or any similar email client that doesn’t let you create folders, set up tags that you can attach to these emails.
That way you can quickly see how many unread Signals you have per each category, making it a lot easier to manage them moving forward.
Now that I use this method I find it a lot easier to manage my Signals, and I find them more useful than ever because they follow a clear structure – they’re giving me relevant information about specific categories that I’ve set up.
Following this example, you’ll soon start identifying patterns and you’ll know what types of emails to expect, while knowing to expect above and beyond what your categories already track.