Small Business Post Intent Analysis
UX Research and Data Science collaborated to develop a better understanding of the different intents Page admins have when they post and how well their posts are helping them meet those intents
Methodology
We sampled 170 posts from each of the 3 top verticals and performed a content analysis on each using a code book generated from an initial small sampling of business posts. The coding involved recording key characteristics of each post (e.g., contains original content, contains visual, any actions requested, etc.) and categorizing the overall intent of a post into one of the following 9 categories:
1. Sales: Online: make online transaction
2. Sales: Physical Location: make offline transaction (including booking an appointment for services, buying tickets to an event, dine at a restaurant, etc.)
3. Awareness: brand building, including building recognition for products/services offered by the business
4. Informational: give information related to the business but not produced by the business (e.g., a chiropractor reshares a meme of the best stretching techniques)
5. Event Invite: invite to an event hosted by the business
6. Business Update: update about the business (e.g., new hours, closing early today)
7. Recruit: ask people to join, volunteer, sign up, apply, etc.
8. Entertainment: information not related to the business (e.g., funny meme, interesting article)
9. Community Build: prompt users to respond and engage with a post or share content (e.g., “Share your holiday drink of choice!”)
Select Main Findings
• We found that the most common admin intent is Awareness, with 1.6x as many posts being categorized as having this intent than the next category, Sales: Physical Location. The next most common intents are Informational and Business Update.
• Posts with Sales intent were the most engaging, with intent of online sales being more engaging than sales at a physical location, followed by Community Build and Awareness, respectively.
What makes sales posts so engaging?
Sales posts' high engagement efficiency could be related to the high percentage of original content as well as usage of visuals.
• We've seen in previous analyses that original content (content created by admins themselves) is more engaging than unoriginal content. Our analysis of admin post intents also showed that this is usually true.
• We also find that including visuals in a post may be related to higher social engagement. The two sales intent categories have the highest percentage of posts that contain visuals of their products or services. We also see that comparing sales posts with and without visuals, there is a large difference in engagement.
• Sales posts are unique in the way that they necessarily have high reliance on both visuals and text to describe a product or service. Text descriptions are especially important for sales intent as admins may need to convey information such as: value proposition, pricing, transaction instruction, etc. (see example posts above)
• Among all intents analyzed, sales posts are among the highest in terms of median text length.
Recommendations
How can we drive more engaging sales posts?
Based on our analysis, some product recommendations for driving more engaging sales posts are:
• Push admins to share their own photos or videos of the product/service they are offering when they have sales intent. This can be included as part of NUX, composer tool tips, and/or the insights tab. In terms of Shops for Pages, this also justifies pushing admins to share listings from their Shops or upload their inventory there.
• Introduce structured posts to help admins with sales intent to better display and organize information for these posts that have high reliance on both texts and visuals. We can also surface this type of structured post via interception as we develop the sales intent model.
• It is critical to ensure that we accommodate these information-heavy sales posts in the new sharing format of the stories-first world. The current stories format favors heavy visuals and less text. An examples of what we can do is consider designing visual elements to replace or redirect common text descriptions (e.g., price, transaction instruction, etc.).
• Train a model to detect sales post intent to enable personalized and structured post interception capabilities.
Examples of posts with Sales intent are shown above this section and below. Note that Posts were categorized in terms of the admin's main overall intent, even though a single post could (and often does) have multiple intents.
