How is AI Enabled Consumer Intelligence (AICI) different from conventional Social Listening, and how can brands set their AICI projects up for success from the very beginning? These questions were tackled on our recent webinar "Social Listening Evolved" with Dr Jillian Ney of the Social Intelligence Lab, and Jenny Force, our VP of Marketing.
We can think of Consumer Intelligence as an evolution of Social Listening. When Social Listening platforms first appeared, around 2007, they gave us instant access to millions of conversations taking place in public social media channels, and that was a powerful new way of understanding consumers. These were unprompted, spontaneous discussions, where people tend to be more open and honest than they would be in something like a focus group or market research questionnaire or study.
The other big advantage was that this data was available in close to real time, so it was possible to get a live view of how consumers were discussing your brand, or any other topic for that matter.
This was all a giant step forward, but it has limitations. Conventional social listening drops a lot of unstructured data onto the user and leaves them to figure out what it all means. If you’ve got skilled analysts available, that’s fine, they can work with that data to find meaning in it, but if you don’t have people with the right skillset, it’s very easy to misinterpret what the data is telling you, or simply overlook a lot of insights.
And this is where AI Enabled Consumer Intelligence comes in. With a blend of modern AI technology, and human expertise in data-science and market research, these new platforms can structure huge volumes of data intelligently and give the user much more help in interpreting what the data actually means so they can find real insights that have business value.
Let’s use a simple example to illustrate the difference. Imagine the marketing team at a Scotch whisky distillery want to build a better understanding of when and why people think about buying their products:
- Social Listening simply tells you that from October to February people talk more about Scotch whisky on social media, and maybe brand X is more popular than brand Y.
- AI Enabled Consumer Intelligence tells you that people enjoy whisky for its warming quality during the cold winter months, and that they’re also interested in finding new whisky cocktail recipes. Furthermore, in the run-up to Christmas, people need help choosing the best brand as a gift for the Scotch aficionado in their life.
Social listening is good at showing you raw quantitative data, but not what that data means. AICI helps you interpret it all accurately, so you can make better decisions based on meaningful insights.
AICI offers a lot of value over standard social listening, but in order for it to work well you have to put a lot more thought into how any research project is set up before you start looking for insights. While social listening tools allow you to simply dive in and start exploring the raw data, AICI needs a more thoughtful approach.
Here are five tips that will help you to set up AI Enabled Consumer Intelligence projects for the best possible chance of success.
1) Understand Your Project's Background
This seems obvious, but understanding the context of what you want to research is critically important. It’s much more difficult to find meaningful insights in social data if the people doing the research don’t understand the context and nuances of the market or subject they’re investigating.
At Linkfluence our team includes specialists with deep knowledge of numerous industries, so when they help clients build AI Enabled Consumer Intelligence projects, they are able to do so with a solid understanding of the unique complexities of those markets.2) Know What Question You’re Trying to Answer
Often businesses embark on Consumer Intelligence projects with no clear idea of what they’re trying to achieve, beyond a ‘fishing expedition’ to see if looking at social data might throw up any interesting information. This is counter-productive and unlikely to yield valuable results.
It’s better to start these projects with a well-defined problem that you want to solve. This will help the team set everything up properly, in a way that’s much more likely to give you some useful answers. Here are some examples of common use-cases, and how you can think of them as questions to be answered:
- Reputation Monitoring - how do people think of our brand and how is that changing?
- Competitive Menchmarking - how do consumers view our brand against our competitors?
- Content Ideation - what topics interest my audience, what questions do they have?
- Influencer Discovery and Vetting - which influencers are most respected amongst my target audience?
- Campaign Performance Measurement - how have my online and offline campaigns performed? How can I measure them all on a level playing field?
- Brand Equity Tracking - is the value of our brand increasing, decreasing, or stagnant - what’s causing that trend?
- Tribes Identification and Activation - who are the main audience segments that are relevant to us, and how do we engage with them effectively?
- CX Analysis - what kind experience do customers have at our retail outlets, which are performing well, which need to be improved? Which of our products are people having good/bad experiences with?
- Trend Forecasting - what’s changing in our market, what new products and experiences do we need to offer our customers to stay relevant and competitive?
But remember that specificity is powerful. The more focused your question is, the more accurate the answer you’re likely to get.3) Identify Which Data Will be Useful
We’re very proud of the fact that we offer the most comprehensive social-data coverage on the market, but remember that not every project requires every data source. One of the advantages of having human research expertise involved in the process is that they will be able to call upon their experience and knowledge of an industry to decide which data sources are most likely to yield the best results in any given context.
4) Structure Data Properly
A mountain of raw, unstructured social data isn’t much use to anybody. You could perform some simple quantitative analysis to get an understanding of trends and keyword volumes, but it’s going to be difficult to uncover meaningful insights.
Let’s use these two tweets as an example - they’re both returned by a search on the keyword “Hamilton” and they’re both about very different topics; the Broadway musical, Hamilton, and the Formula One racing driver, Lewis Hamilton.
With unstructured data, you’ve got this kind of confusion happening at a huge scale and while it’s possible to use well constructed boolean searches to limit the problem, it’s very difficult to completely eliminate it. So the data is inevitably going to be ‘noisy’ to varying degrees.
AI Enabled Consumer Intelligence tools, such as Linkfluence, structure the data intelligently. For example, we classify the data to three levels of detail, based on topic categories that are widely used across.the digital marketing industry. In this example, there would be no risk of false positives in our data, because the platform has classified all of the data points (whether they’re tweets, comments, reviews, or anything else) and can easily differentiate between Hamilton the musical and Hamilton the sportsman.
We can even do this with images, since modern AI is easily capable of understanding the contents of an image file and classifying it in the same way as textual content.
5) Interpret and Socialize Findings
It’s important to understand that findings and insights are different things. Uncovering patterns and trends in social data is only the first part of the puzzle. It takes human expertise in market research methodologies and experience of the context and nuances of the relevant industries or markets to interpret those findings into accurate, actionable insights.
Without that relevant expertise or domain knowledge it’s all too easy to jump to the wrong conclusions and make bad decisions.
It’s equally important to socialize the insights to relevant stakeholders across the business in a way that leaves little ambiguity or room for misinterpretation. The whole point of doing all this work to uncover consumer insights is to create value for your business, and that can only happen if the insights are communicated clearly to the people who are able to use them to make more effective decisions.
If you're thinking of moving beyond social listening to AI Enabled Consumer Insights, a good next step would be to take a few moments to participate in our Consumer Intelligence Maturity Model.
Simply answer a few questions about how your business already uses social data, and the model will let you know how you compare against other organizations and what steps you can take to become more advanced.