Will social intelligence replace surveys and focus groups in the next five years? Probably not.
But should it be in every market researcher’s toolkit? Absolutely.
Microsoft is using social intelligence to inform marketing and engineering decisions, not just tactical social media implementation. Using public social media data from Twitter, Reddit, Instagram, blogs, forums and reviews, we build insights on customers and the competitive landscape around brand equity, product naming, positioning and messaging, unmet needs, and category trends.
We’re not just looking at how many likes our post got, but questions such as:
- What do customers really think about our products and our competitors’ products?
- Which features do they love, and which ones not so much? What features should we prioritize building?
- What’s new in our category, and do customers actually care about it?
We have gleaned key findings not just on the broad consumer and gaming markets, where you would expect to find an abundance of discussion, but also on the commercial market and across niche audiences like software developers, IT professionals, and educators.
Here are a few of the reasons Microsoft found it useful:
- It’s faster and often far less expensive – so even when there isn’t time or budget for a survey or focus groups, we can still give our stakeholders insight into what’s going on in the market.
- It’s in the customers’ own words, giving us direct access to an unprompted and unfiltered view of how customers perceive our brand and products.
- It’s qualitative data at the scale of quantitative – in fact, the challenge with social data is not so much getting enough data as paring it down.
- It’s real-time, so we can identify category trends as they are emerging and quickly understand customer reactions to our brands and products.
So how do you set up a social intelligence framework that can deliver robust findings? And how do you develop ways to analyse social data to address new and different types of business questions?
By constructing robust queries, getting intimate with social data, and innovating ways to analyse it, companies can build a social intelligence practice that can complement existing market research methods, and can step in when other methods can’t do the job because they require too much time or budget.
Leveraging social media data for market research involves:
- Constructing robust queries to capture the right data
- Queries are to social intelligence what questionnaires are to survey research – these are Boolean queries that define both what you include in your dataset and what you exclude (e.g., include ‘Word’ the word processor, but exclude ‘word’ the word – this can be complicated!). Operators like OR, AND, NOT and parameters like exact match or proximity helps refine the output of the search.
- Some of our queries are really simple, and others contain thousands of terms to make sure we’re getting the right dataset (e.g., whereas the query for ‘Microsoft’ is fairly straightforward, the query for ‘Apple’ is not)
- Our approach is to build unidimensional queries we can combine and re-combine for different projects rather than having to build a custom query for each project – so we have a query for ‘security’ that we can use with ‘Windows’ or ‘Mac’ or ‘Chrome’
- We work with cross-functional teams to capture all relevant terms to include and establish the boundaries of the topic we want to capture (e.g., for the purposes of your business question, do you want to consider machine learning a subcategory of artificial intelligence or a separate topic?)
- We use an iterative process of pulling sample social data to identify additional terms that people use for the topic we’re trying to capture, along with co-mention requirements and exclusions to get a clean dataset (excluding spam, adult content, etc.)
- Understanding the social media universe to correctly analyse and interpret data
- Why people post on social media, the dynamics of social media, and what gets amplified – this will vary by social platform and impact the things people talk about (e.g., people talk about new news rather than old news, their passions and personal lives more than their work lives) and how they talk about them (whether the social currency on the platform encourages people to laud or troll).
- How people use different social media platforms – what kinds of content get posted on Twitter vs. Instagram vs. Reddit and leveraging, and interpreting the content with that in mind; forums can be a wonderful place to get a read on what software developers are discussing – they’re often using them to troubleshoot an issue.
- How to find the right audiences and conversations – whereas in surveys you can simply screen respondents to get the right audience, with social data you have to get a little crafty. You can use a curated list of social handles or leverage people’s self-identification in their bios or posts. And then there are certain forums where the people who are engaged are the audience you’re looking for (e.g., software developer forums) and some topics are only discussed by the audience you’re interested in (e.g., ‘Kubernetes’ – if you’re talking about it, you’re probably the audience we’re looking for).
- Designing social analysis to address strategic marketing questions, e.g., product positioning and messaging, naming, feature development, unmet needs, brand equity, competitive analysis, and category trends:
- Does relevant content exist? Where? What is the context? How can we isolate?
- We developed a methodology for naming to assess relevance, sentiment, context, semantic range, and brand association for product names under consideration.
- We also developed a competitive product analysis to identify features/benefits that matter to customers across the competitive set to inform product development and positioning.
- More recently, we developed a social satisfaction methodology to understand what customers like and dislike about our products and features through review data.
It can take a bit of effort to put the foundations for social intelligence in place, but once you’ve set them up, they can enable your market research team to be very nimble in providing business guidance in today’s rapidly changing world.
