This is the final post in a 3-part series that addresses the future of real-time, multi-channel marketing and the critical, driving role that social media plays in it.
In our first and second posts, we discussed how real-time social data can make an enterprise more agile and responsive, and we discussed the critical role that marketing automation platforms play in making this possible.
In this post, we’ll show you how to get access to real-time social data, extract meaningful signals from it and leverage them in your marketing automation platform to trigger emails, score leads and improve marketing performance.
Social Triggers: Like Any Other (but different)
Most marketing automation platforms can leverage online behaviors to trigger a variety of marketing actions today. For example, many will let you automatically send an email to a contact when they visit a certain page, or launch a drip campaign after a successful lead capture.
Social signals of intent are another class of behavioral trigger that many leading MAPs can leverage today (with IT assistance) to trigger emails, send internal alerts, launch campaigns, score leads – pretty much anything your marketing automation system can do.
By “social signals of intent”, I mean social media users who are asking questions about, complaining about and explicitly buying your type of products and services. These are just three examples – there are actually dozens of types of social signals of intent that can be leveraged to trigger marketing actions.
Using social signals to automatically trigger marketing actions offers several advantages over web- and other behavioral triggers, including:
- more behavioral triggers per contact: your socially-active contacts are publishing a huge amount of data about their likes, interests and feelings in social media. More signals means you have a chance to interact more often, around very specific topics. This shortens sales cycles and keeps your prospect engaged with you.
- higher-quality signals of intent: a traditional marketing trigger is really just a guess about someone’s intentions based on an action like downloading a white paper or clicking on a certain page. A person’s words provide much more information about their intent. In fact, some social signals are extremely explicit requests to buy!
- eliminate the ‘creepy’ factor in social media marketing: if you identify a person complaining about a competitor on Twitter and then send them a tweet, it can be kinda creepy. But if you send them an email pointing out the differences between your products and your competitors, the “stalker risk” is basically eliminated from the equation.
This article explains how to build social triggers you can use in your marketing automation system, in four steps.
Note: you do not need to be active in social media in order to to use socially-triggered marketing, because it does not require a response in social media.
Step 1: Get Real-Time Social Data For Your Business
The first step, of course, is to get real-time social data for your business.
- By real time, I mean pulling new data in every hour or so.
- By social data, I mean grabbing posts published on the social networks, forums and blogs your contacts & prospects use the most.
- By for your business, I mean gathering posts published by your existing network of prospects, customers, followers and contacts PLUS gathering public posts containing keywords related to your business.
There are several places you can obtain real-time social data for your business, including:
from the social networks themselves, using their Open APIs.
from your existing social media monitoring platform, such as Salesforce Marketing Cloud or Oracle Social Relationship Management.
Depending upon your business objective, the data you acquire may be for your followers only (Facebook fans & Twitter followers); all public posts about topics that matter to your business (Twitter, blog comments, forum posts); or, both.
Step 2: Mine Your Social Data for Signals That Matter
Next, you will need to mine your social data for suitable trigger signals. In our experience, about 3-5% of all tweets contain some form of commercial intent (about 300 million events per month on Twitter / English language alone).
Your options for identifying trigger signals in your social data include:
Manually classify your social data (ie, monitor and flag for action). This is the most common and most precise method available, but there are serious drawbacks. First, you can only scale this method by adding people. Second, your recall (market coverage) will suffer whenever your labor is tapped out or off-duty – meaning you will miss some opportunities. Third, you may struggle with the timeliness and consistency of your results, which means you will miss more “windows of opportunity”.
Machine-classify using keywords and/or text analytics: Some companies use a combination of natural language processing (text analytics), Boolean keyword expressions and/or sentiment classification to approximate “intent”. This method can be automated, which saves time. Unfortunately, to get results anywhere close to useful, you will either need to become a boolean keyword ninja or employ expensive linguistics experts to build natural language filters for your business. Otherwise, the precision and recall of text-mining methods will not be very good. This means you will get a lot of signals that aren’t what you’re looking for, and you will miss a lot of opportunities. Most customers we’ve worked with who have tried keyword-mining approaches have been unhappy with the quality of results.
Machine-classify using a specialized intent-mining platform like NeedTagger. Intent-mining platforms offer better results than keyword and text-mining approaches in terms of cost, precision and recall – but they are still not as precise as human beings. Some platforms like ours leverage more data than text to classify intent, which is what creates the precision advantage. NeedTagger also lets you create your own ‘intent filters’ for any need and test them in real-time against Twitter. Other vendors that offer intent-classification services include LeadSift, Solariat and Aiaioo.
Machine Detection of Intent: NeedTagger’s Social Signals API
We’ve developed an API that mines your social data for signals of intent. It works with any short-form social posts, in real time.
Out of the box, our API identifies more than 70 signals of intent, which are grouped into four basic types:
14 conversation types: Asking Questions, Buying, Wishing, Complaining, etc.
buying signals for over 60 common product & service types: buying a car, getting a mortgage, applying for college, etc.
Life Events: moving, recent pregnancy, a new job, etc.
In addition, you can build custom social signals for your business in minutes using our self-service app (no other provider offers this flexibility).
Every post submitted to our API is analyzed for the four types above. Results are provided in seconds.
Our API is optimized for Twitter, but it classifies any short-form social media like the kind you find on Twitter, Facebook, blog comments and more. Right now, we are limited to the English language.
Create Custom Social Signals For Your Business
Using our Social Signals API, you can build custom “social signals” for any type of need or requirement and immediately gain access to them from our API.
The diagram below explains how this works:
Creating a new social signal for your business is easy. Our StreamBuilder tool lets you create a custom signal filter and view the types of posts it finds, in real time. StreamBuilder works a lot like Twitter Advanced Search, just with more search options.
If you’d like to learn more about our new Social Signals API, use our API Inquiry Form or send an email to info@needtagger dot com.
Step 3: Get Social IDs for Your Marketing Contacts
In order to take full advantage of the signals you receive, you’ll need to know the social IDs for the contacts in your database, so when a social signal comes in you will know who posted it.
Most MAPs today allow you to maintain social ids for contacts. If you are doing this religiously, then move on to Step 4.
If you aren’t maintaining social IDs for your contacts, you can leverage third party services such as Social123, Fliptop and FullContact to maintain the social data for your contacts, automatically.
Couple of cautions, though:
no provider can map social IDs to email addresses with 100% coverage.
the mapping process works 1-way: you can retrieve social account IDs for email addresses & social IDs, but you won’t be able to retrieve an email address for a given social ID.
To keep your contact database fresh you will need to send all new contacts through your provider’s API immediately, and you may want to build a polling routine to update existing contacts every month or so.
Step 4: Create Social Triggers In Your Marketing Automation Platform
The final step is to create social triggers in your marketing automation platform that can be used to send emails, score leads or whatever you think is appropriate.
Some MAPs are better than others in utilizing external data to trigger events. Silverpop and Marketo are two that can do this today (with IT assistance).
The process of building social triggers is the same used to create any other type of trigger, except the trigger signal is from social media. Each MAP has its own process for creating a marketing trigger, so I won’t go into that in detail.
From this point forward, your social triggers will work like any other trigger – email, web hook, landing page, whatever.
The “social triggers” discussed above can be implemented today with some packaged marketing automation platforms. If you use an in-house marketing automation system, you will need IT’s assistance, but it can be implemented today.
We hope this article has opened your eyes to the new marketing automation capabilities that real-time, social data can provide for your business.