Use cases

Sort every mention: crisis, complaint, lead, praise, spam

Reads mentions and comments for safety reports, fake support accounts, complaints and leads. It suggests hiding spam for a moderator, never complaints.

Try it on this example

Example · A comment under the launch post: a wheel lock that failed, and a fake care account asking for a card number

Source (own-page comment, public mention, review site or direct message): Own-page comment (Instagram)

The post it replies to (empty if none): Meet the Glide 2: our lightest stroller yet, and it folds with one hand. Now in stores and online.

The mention or comment

The wheel lock on our Glide 2 snapped on the second day and it rolled off the curb towards the road with my son in it. Thank god a man walking past grabbed it. I sent you a message about it and now a "Little Otter Care Team" account has DMd me asking for my card number and the code on the back to "process a refund". Is that even you?? Nobody from the real team has answered me in three days.
  1. Is the post about Little Otter or its products, and not about something else with a similar name?Yes97%
  2. What kind of post is this, mainly?Crisis signal100%
  3. Does the post describe an injury, an illness, a fire, a danger to a child, or a product that could hurt someone?Yes98%
  4. Does the post show or describe someone posing as Little Otter or its staff, or a fake offer in our name?Yes98%
  5. Would a careful brand reply to this post, in public or by direct message?Yes96%
  6. Does resolving the post need order, account or personal details, so the conversation should move to a direct message?Yes73%
  7. On our own page, should a moderator keep this comment visible or consider hiding it?Keep visible100%
  8. Does the author say they are a journalist, a creator or a partner of the brand?No92%
  9. How fast should the brand act on this post?Critical100%

These are real answers stored from one run on this example.

The prism behind it

Sort every mention: crisis, complaint, lead, praise, spam9 questions

Fields

  • Source (own-page comment, public mention, review site or direct message)
  • The post it replies to (empty if none)
  • The mention or comment

Context

These are social media mentions, comments and messages about Little Otter, a brand of strollers, car seats and nursery furniture sold online and in stores. Each item arrives from our social listening tool. The source field says whether it is a comment on one of our own posts, a public mention elsewhere, a post on a review site or a direct message to us. The parent field holds the post a comment replies to. Our brand and products: Little Otter; the Glide and Glide 2 strollers, the Anchor car seat, the Nest crib and the Hop high chair. "Otter" on its own is also an animal and a road name, and "Glide" is used by other brands. Our only official accounts: @littleotter on every platform, and @littleottercare for customer care. We never ask for a card number, a password or a payment by direct message or comment, and we never run a giveaway that asks for payment details. Current campaign: the Glide 2 launch. Known incident: none open. How we act on the answers: a safety report or a crisis alerts the on-call social lead and our product safety team. Accounts posing as us are reported to the platform by a person, and followers are warned. Complaints go to care, and buying questions to the sales team. On our own posts, a moderator may hide spam, scams, hate, and posts that expose someone else's personal details. We never hide a complaint, criticism or a negative opinion, however strong: hiding genuine criticism misleads other customers. Follower counts, reach and likely spread come from the platform data in our code, not from these answers.

Questions

  1. Is the post about Little Otter or its products, and not about something else with a similar name? Yes / No

    Use the brand and product names in the context. A post about the animal, a road, or another brand's Glide is not about us. A comment on one of our own posts counts as about us when it responds to the post or to us. Yes: The post is about our brand, our products, our staff or our accounts. No: The post uses a similar name for something else, or is not about us at all.

  2. What kind of post is this, mainly? Choice

    Pick the one option that fits the post as a whole. When a post contains a report of injury, danger or a serious accusation, pick Crisis signal even if it is also a complaint. When a post both praises and complains, pick Complaint.

    • Crisis signal Reports an injury, an illness, a fire, a danger to a child or a product that could hurt someone, threatens legal action, or makes a serious accusation such as fraud, discrimination or a data leak.
    • Complaint A customer problem with a product, an order, a delivery or our service, with no injury, danger or serious accusation.
    • Buying question Asks about buying, price, stock, stores or which product to choose, with no complaint.
    • Other question Asks for information such as how to use, clean or fold a product, with no complaint and no buying intent.
    • Praise Describes a good experience or recommends us, with no complaint.
    • Spam or scam Unrelated promotion, a link farm, a fake giveaway, or an account posing as us or as our staff.
    • Abuse Harassment, threats or hate aimed at a person or a group, with no customer problem behind it.
    • Passing mention Mentions us in passing, jokes or tags a friend, with nothing to answer or act on.
  3. Does the post describe an injury, an illness, a fire, a danger to a child, or a product that could hurt someone? Yes / No

    Count what the post says happened or nearly happened, and a product fault that could cause harm, such as a brake that fails, a strap that snaps or a part a child could swallow. A joke or a figure of speech does not count. Yes: The post describes harm, a near miss, or a fault that could hurt someone. No: The post describes no harm and no dangerous fault.

  4. Does the post show or describe someone posing as Little Otter or its staff, or a fake offer in our name? Yes / No

    Compare any account named or described with the official accounts in the context. Count an account that is not official but presents itself as us or as our care team, a message asking for card details, passwords or payment in our name, and a giveaway or discount we do not run. The post itself may be from the scammer or from a customer reporting one. Yes: The post is, shows or describes an impersonation or a fake offer in our name. No: Nothing in the post suggests anyone is posing as us.

  5. Would a careful brand reply to this post, in public or by direct message? Yes / No

    Judge whether a reply from us would help the author or the people reading, such as answering a question, taking up a problem, warning of a scam or thanking someone for detailed praise. Yes: A reply from us is called for. No: No reply is needed, as for spam, abuse or a passing mention.

  6. Does resolving the post need order, account or personal details, so the conversation should move to a direct message? Yes / No

    Count problems that need an order number, an address, a photo of a fault, or other details the author should not post in public. Yes: Resolving it needs such details. No: It can be answered in public, or needs no answer.

  7. On our own page, should a moderator keep this comment visible or consider hiding it? Choice

    Use the source field. Our policy in the context allows hiding only spam, scams, hate, and posts that expose someone else's personal details. A complaint, criticism or negative opinion is always kept visible, however strong, even when it also mentions a scam.

    • Not on our page The source is a public mention elsewhere, a review site or a direct message, so there is nothing for us to hide.
    • Keep visible A comment on our own post that is not spam, a scam, hate or someone else's personal details. Every complaint and criticism belongs here.
    • Suggest hiding A comment on our own post that is only spam, a scam link, hate, or exposes someone else's personal details, with no complaint in it. A moderator decides.
  8. Does the author say they are a journalist, a creator or a partner of the brand? Yes / No

    Go only by what the post says, such as "writing a piece on", "for my channel" or "as a brand ambassador". Follower counts and verified badges are checked by code. Yes: The author says they are a journalist, a creator or a partner. No: The author says nothing of the kind.

  9. How fast should the brand act on this post? Scale

    Judge from the post alone what waiting would cost the author, other customers or the brand. Pick one level.

    • Low Nothing to act on, or it can wait several days, such as praise or chatter.
    • Normal Should be answered within a business day, such as an ordinary complaint or question.
    • High Should be answered within hours, such as an angry complaint about an order that is due or a problem others are repeating.
    • Critical Should go to PR, product safety or legal now, such as a report of harm to a child, an impersonator targeting customers, or a serious accusation.

Lens columns

about_us, about_us_probability, kind, kind_probability, safety_issue, safety_issue_probability, impersonation_or_scam, impersonation_or_scam_probability, needs_reply, needs_reply_probability, move_to_private, move_to_private_probability, own_page_moderation, own_page_moderation_probability, press_or_creator, press_or_creator_probability, urgency, urgency_average

Run it on your own text

Add this prism in the app, change any question, and test it on a file of your own.

Ask for an invite