The myths and realities of social intelligence and analytics

At the start of 2022, Synthesio deemed this year the “year of AICI.” In many ways, this was true: we saw an increasing number of clients embrace combining online and offline data sources (like social + survey or search), and social listening “grew up” even more. The value of social data continued to move beyond social marketing teams to empower insights, brand, and even innovation pros. And companies allocated more budget to social data; The Social Intelligence Lab reported in its State of Social Listening 2022 report that 33.5% are spending $100K or more on social data tools every year (up from 10% in 2019).

But despite social intelligence’s growth, we still hear a lot of misconceptions about what it is and how best to use it. So to start 2023 properly, we’d like to dispel some of the most commonly occurring myths:

Myth 1: Social media data is worthless – it is all Twitter data and it doesn’t tell me anything

Reality: Twitter data can sometimes dominate results, if you allow it to. And sometimes it’s very useful for brands; even the microblogging, short and sweet text has value in it, especially in its volume. But Twitter data isn’t right for every single question or approach. Just as you would consider how to build a good sample in a qualitative or quantitative research project, it is important to consider your “data universe” when gathering social media data. You need to access a wide variety of sources in your chosen platform before launching a social media research program. On several occasions, specialized blogs and forums are more insightful than Twitter (for example, healthcare/pharma companies might get more out of Reddit health communities or dedicated forums like MedHelp). It is also critical to treat the data coming from various sources as a heterogenous mass. A tweet about “Tesla” for instance will carry a very different meaning layer compared to a “Tesla” usage experience review or a “Tesla” post in a forum on the future of mobility. This is where experienced analysts come in: they can bring this data to life, pretty much in the same way we’d split and report our survey data comparing target groups, socioeconomic or attitudinal criteria.

Myth 2: Social intelligence is only for understanding our PR and social media marketing efforts and to manage crises

Reality: Social media data and intelligence can be useful in those cases, but there are so many more questions that it can help with. We have used it reliably in many cases, including to guide and supplement segmentation and audience understanding; to understand a market landscape and broader context; to understand signals of macro trends and deep dive into micro- and nano-trends; to surface insight about why people engage (or not) with particular products and services; to understand specific brand moments that comprise a brand experience; and to get a clear picture of what makes people tick and what ticks them off on a wide range of topics. This is by no means an exhaustive list, but these are approaches we’ve taken time and again to create actionable insights.

Myth 3: We can just look for what’s interesting within all social media data

Reality: Given enough time and money, you might eventually get there (while also burning an incredible amount of electricity), but resources are rarely unlimited. As the saying goes: “If one does not know to which port one is sailing, no wind is favorable.” Therefore, it is critical to create some boundaries around your research, so that you know you’ve found something interesting. Your human expertise and industry knowledge will help you to develop some relevant boundaries.

Myth 4: No one talks about this topic and, therefore, I can’t do any research using social intelligence

Reality: Sometimes, we can be a little too specific in our requirements. This is not uncommon with brand research. It does not mean that there is no value that social media data can provide – in fact it can help steep us more in the consumer reality. Consumers might not talk about a specific brand’s product – for example a specific brand of frozen peas or tinned tomatoes – but they will talk about very relevant topics, such as how they cook at home, favourite flavours and dishes, and what they aspire to bring to their mealtimes.

Myth 5: Social intelligence will be fast and cheap

Reality: It can be faster and cheaper than more traditional methodologies for sure. It can also give you deeper insight into what is moving people because it is unprompted. And advances in AI, like our Topic Modeling, are making the path from data to insights even shorter. But because there are layers of refinement needed to prepare a strong and relevant dataset as well as human-led insight, good social intelligence and analytics requires a certain level of investment.

Myth 6: AI does it all

Reality: Analysis of any unstructured data would be incredibly restricted without AI. In the first place, it would be nearly impossible to find the data we wanted to review. The power we have as human beings is to be able to find meaning and importance in the factors that AI brings to our attention. AI is powerful and can show us patterns we might not have noticed ourselves. But it takes a human being to say why that is important, assuming it is a human being who understands the question and the broader category and topic. Setting expectations around AI is a key tip we heard from our customers in our panel on scaling social intelligence.

Myth 7: All tools are created equal

Reality: Each tool has its pros and cons and there are always trade-offs depending on what you want to do. There are many different factors to consider, depending on what’s most important to you and your organization – data sourcing, level of cleaning, ease of use, built-in widgets, volume of data, level of in-built analytics and AI, language and geographic coverage, etc. What’s more, these are changing all of the time. It can feel overwhelming to stay on top of it, but regular reviews of needs versus what you can get are important.

Myth 8: It is very easy to become an expert in social intelligence

Reality: Like any research approach, a little knowledge can be dangerous. You cannot become an expert in all aspects of social intelligence and analytics. But it is possible for a keen and curious mind to learn how to make the most of unstructured data and how to execute worthwhile and reliable social media data research and insight.

It’s also increasingly accessible (and important) for non-experts to use insights from social data in their daily jobs. Some of our clients have set up self-service dashboards for their peers to view key metrics, or do a quick check on what consumers are saying about a particular topic. Others have linked social listening KPIs to their sales metrics and other business KPIs.

But social intelligence requires research rigor (especially when analyzing multiple data sources or applying predictive models!) which is why many of our clients opt for a hybrid approach, combining SaaS technology with expert services.

This article was originally posted on the Synthesio website.

‍