Sprout Social Listening: Topics, Sources, Limits and the Add-on

What Sprout Social Listening is, and how you get it

Sprout Social Listening watches public conversation for the words you care about: your brand, your products, your competitors, a campaign hashtag or a whole industry. Each search you set up is a Topic, a saved query that keeps collecting messages for as long as it stays active. Inside a Topic you get volume, engagement, sentiment, recurring themes and the messages themselves. It is a different job from the Smart Inbox. The inbox handles messages sent to your own profiles. Listening finds the ones that never were.

Listening is a paid add-on, and Sprout publishes no price for it. It is not part of Standard, Professional or Advanced by default, and it can be added to Standard and up, never to Essentials ^[1]. Only the Account Owner or a user with the Manage Billing permission can buy it, through an account manager or sales, and the number of Topics you get is written into that agreement. Extra Topics are bought the same way, again with no listed price ^[2]. Plan on a sales call.

Two practical consequences follow. The 30-day trial, which needs no card, does not include Listening; Sprout points trial users to a demo instead. Without the add-on you can still preview a query, but you cannot save it as a Topic. Listening also sits on top of a base plan priced per seat. All annual plans renew automatically: annual plans paid upfront in the app are cancelled in the billing page before the renewal date, while annual plans paid monthly and invoiced plans need 30 days' notice by email. Removing an add-on brings no refund. For the full plan math, see our Sprout Social pricing guide.

Setting up your first Topic

Every Topic starts from one of five templates: Brand Health, Industry Insights, Competitive Analysis, Campaign Analysis and Event Monitoring. Pick the one closest to the decision you want to inform; Brand Health is the natural first Topic for most teams. Build a Campaign Analysis Topic before launch, with the campaign hashtags already in the query, because a Topic only collects forward from its start plus a limited backfill.

Your Topic allowance is shared across the whole account, not counted per Group, so one busy brand team can use up slots another team was counting on. Archived Topics do not count toward the limit, and neither do Featured Topics, which are curated queries Sprout maintains and you cannot edit. Access is controlled through feature permissions: Manage Topics lets a user build and edit, while view access lets people read results. Anyone with view access can see every Listening report in the Groups they belong to, which is worth knowing before you add a sensitive competitor Topic.

Before you commit, use the preview. It samples X, Reddit and Bluesky over the past 30 days and now returns in three to five seconds. It leaves out reposts, accounts with zero followers and explicit content, so treat it as a sense check on volume and noise, not a forecast. The other sources only start flowing once you click Start Listening and the Topic is saved.

Writing the query without Boolean

You do not need Boolean skills. The query builder takes keywords in plain fields and joins them with AND and OR logic behind the scenes, and you can group terms so that results must mention all of them together. Keywords ignore case and accent marks. Exact Match switches off fuzzy matching, plurals and conjugations when a brand name collides with an ordinary word. Proximity Match, spam keyword exclusion and an explicit content filter sit in the same builder, and an AI spam filter, tuned for English, is added to every Topic automatically.

There is one hard limit: a query can hold at most 2,037 characters, on every plan. That sounds generous until you list a dozen product names, misspellings and exclusions. Keep the core query tight and move the detail into Themes, which are saved keyword sets used as filters inside a Topic. A Topic can hold up to 50 Themes, each named in 30 characters or fewer, and a message joins a Theme only when it matches the keywords exactly.

AI helps at two levels. Keyword suggestions run on OpenAI through the OpenAI integration, which is on by default; an Account Owner can turn it off. The newer AI query builder, launched with Trellis in September 2026, turns a plain-language prompt into a full query. It is a beta open to every plan with the Listening add-on and covers every single-query Topic type except Competitive Analysis. Only the prompt you type goes to the model, per Sprout. Edits now reprocess most data in under a minute, though Theme updates and full linguistic matching still run on a 12-hour batch.

Which sources it covers, and where they fall short

Listening reads Facebook, X, Instagram, LinkedIn mentions, Reddit, YouTube, Tumblr and the web, and it has since added Bluesky and TikTok. That list hides a lot of fine print, and the fine print decides what your charts mean. X is the richest source, covering posts, replies, mentions, quotes and reposts, with new posts arriving in two to three minutes. Facebook, Instagram, YouTube and Tumblr are checked about every four hours. The rest is uneven:

• Facebook covers Pages only. You name the Pages, up to 250 per Topic.
• Instagram is found by hashtag, so a post that only mentions you in the caption text may be missed. It needs an Instagram Business Profile attached to your Sprout Group, Instagram allows 30 unique hashtags per Business Profile every 7 days (connecting another profile adds 30 more, and collection stops if the Instagram or Facebook token expires), and public Instagram data has no backfill and no comments.
• YouTube gives video descriptions, not comments. Tumblr gives articles, not reblogs.
• LinkedIn cannot be searched as a network; you get comments and mentions on the company pages you have connected.
• Web means blogs, forums and news. Reviews, general web pages and paywalled content are left out, and after a feed change in August 2026 some sites that used to appear may no longer do so.
• Reddit listening is keyword based. Pulling every post and comment from a chosen subreddit needs the separate Advanced Listening add-on.
• Bluesky covers posts, replies, mentions and quotes; a mention only matches the full handle, domain included.
• TikTok cannot be searched by keyword, because TikTok's API exposes only your own content. You get the brand hashtags TikTok links to a connected TikTok business profile, plus comments and mentions on your connected videos. Author information is not available for brand-hashtag data, and posts by underage or unknown-age users are left out.

Pinterest is not a Listening source at all, so pin-heavy brands will see nothing there. Location is thin too: the country filter applies to X data only, and since X removed message-level location from its API in early 2026 it mostly covers older X data; Listening cannot target a geographic area. For mentions on your own profiles, the Connected Profiles option pulls Facebook, Instagram, LinkedIn and TikTok data, but it excludes your brand's own posts and cannot feed a Competitive Analysis Topic.

How far back a new Topic can see

A new Topic backfills 30 days of history automatically on every network except Instagram. Some sources stop sooner: X backfills 30 days or 200k messages, and Reddit and Bluesky 30 days or 100,000 messages. Anything older has to be bought. Sprout sells 12-month and 24-month backfills through a Customer Success representative, as a separate purchase with no published price ^[3].

Timing matters more than it looks. A Topic built the day a crisis starts captures the month before and everything after, but a Topic built a quarter later cannot recreate the peak without a paid backfill. The same applies to Connected Profiles: adding one to an existing Topic loads nothing from before the day it was added. Competitors or keywords added to an existing Topic work the same way: collection starts when the edited Topic is saved, with no backfill. Build the Topics you expect to need before you need them, starting with Brand Health and your top competitors.

Reading sentiment, Themes and Smart Categories

Sentiment tracking and scoring come with the add-on, so a Listening customer on Standard sees sentiment on Topic messages even though sentiment in the Smart Inbox and reviews is an Advanced feature. The two share one ruleset. Writing your own Sentiment Classification Rules, though, is reserved for Advanced and above. When the model gets a message wrong, you can reclassify it, up to 1,000 messages at once from the Message Explorer. The correction applies to that one Topic only, does not change the Smart Inbox and leaves no audit log. Agree as a team who may reclassify, or the trend line will drift with whoever touched it last.

Smart Categories group the nouns in a Topic into Person, Place, Product, Organization, Media, Event and Policy, the quickest way to spot who and what pulls conversation your way. They work in English only and export to CSV. Read the metrics with their gaps in mind. Potential impressions exist only for X, YouTube, Reddit and connected Facebook data. Engagement is missing for Reddit and the web. Unique authors leave out Instagram and the web, and one help page also excludes TikTok. A chart that blends networks can look like a trend when it is really a change in source mix.

Spike Alerts and competitor share of voice

You set Spike Alerts per Topic on volume, potential impressions, positive sentiment or negative sentiment, choose a sensitivity level, and Sprout emails the recipients when the line jumps; alerts also appear in the Smart Inbox. Creating one needs Manage Topics, receiving one needs view access, and every recipient must belong to the same Group as the Topic. A negative-sentiment alert on your Brand Health Topic is the early warning most teams buy Listening for.

For competitors, the Competitive Analysis template compares brands inside one Topic. Share of voice is defined simply: messages matching a competitor divided by all messages in the Topic. The result is only as fair as the query. These Topics cannot use Connected Profiles or the AI query builder, and the repost breakdown charts do not apply to them. They are distinct from the competitor reports on Professional and Advanced, which measure rival profiles' own posts rather than the wider conversation about them.

Trellis in Listening, and the credit meter

Trellis, Sprout's AI agent, reads Listening data when the add-on is on your plan. You ask in plain language and it summarizes a Topic, flags sentiment shifts and emerging themes, and suggests campaign ideas. Sprout documents summaries of up to 1,000 messages, with a production limit of 250 messages per query. Teams without Listening get only the Performance and Engagement agent, so the Listening agent is part of what the add-on unlocks. Trellis Studio, a limited release open only to select customers, also lets skills run on a schedule against a Topic; the Listening sentiment skill looks at the past seven days.

Trellis runs on credits. Every seat gets 100 a month on every plan, and only Trellis chat and skills draw on them; AI Assist features such as Analyze Topics or message summaries do not. Credit use grows with prompt complexity and the volume of data retrieved, so broad questions about a large Topic burn through an allowance fast. Heavy users can buy Trellis Plus, which raises the allowance to 1,000 credits under a fair-use cap and costs $28 per user a month billed annually or $35 billed monthly.

Getting data out: exports, sharing, archives and the API

The message table exports to CSV, capped at 10,000 messages per file, taken from the top of whatever sort order is active. Sort by engagement before you export if you want the loudest posts, or by date if you need a clean slice. Listening Insights can be shared by link with people who have no Sprout login, once the Account Owner turns link sharing on, and a PDF can be emailed to up to 25 recipients. Scheduling that PDF weekly or monthly needs the Advanced plan or Premium Analytics. Premium Analytics is also what puts Listening widgets into the custom report builder.

Archiving a Topic frees a slot, but its data covers only the period it was active, and an archived Instagram hashtag slot is released only after a seven-day decay. Archived Topics go dormant after 60 days unviewed, and active ones after 180 days; waking one takes hours, quoted at two to three in one help article and about six in the FAQ. If an archived Topic still matters for year-on-year comparisons, open it at least once every 60 days so it stays awake.

The Sprout API returns Listening Topics, metrics and messages, but only on Advanced with API permissions. It returns no X Listening data and no message-level Reddit data. Metrics can be bucketed by day, network, sentiment or language, and each message carries a raw sentiment score between -1 and 1. The topics endpoint also reports how far back each Topic has data. Our Sprout Social API guide covers the rest.

Two 2026 changes that rewrite your charts

First, the Demographics tab is gone. X removed message-level location data from its API in early 2026, which left the Listening location widgets nearly empty, and Sprout removed the tab from all Topics on August 3, 2026. The Locations widgets have no replacement, and the Profile Overview moved to the Conversation tab. Sprout calls it an industry-wide API limit. Ignore any marketing page that still promises age, gender or location breakdowns. The help center now says Listening has no demographic data at message level.

Second, reposts changed in October 2026. Listening now rolls X and Bluesky reposts up into metrics on the original post instead of storing each one as a message, and historical standalone X retweet records were permanently purged ^[4]. Repost volume is dated to the original post, so past volumes can shift. When reposts are included, the original's sentiment and keywords count once per repost, which can magnify a single viral post. Quote posts still appear as their own messages. If you compare this quarter with last year, note the break in the series.

When Listening is worth the add-on

Listening earns its price when public conversation drives a decision you would otherwise make blind. A consumer brand with real volume on X, Reddit and the news, a team that needs early crisis warning, or an agency that pitches with share of voice and industry trends will use it every week. It also fits teams already living in Sprout, because alerts land in the Smart Inbox and Listening messages can be answered, liked or sent to Compose without switching tools.

It is a weaker buy for small brands with little organic talk, for teams whose audience talks mostly in public Instagram posts beyond hashtags, on Pinterest or in YouTube comments, and for anyone who needs demographics or long history without paying again for backfill. Essentials users cannot add it at all. Before you sign, ask sales four things: how many Topics the price covers, what a 12-month backfill would add, whether you need the Advanced Listening add-on for subreddits, and how the add-on sits in an annual contract with no refunds. For a rival's approach, see our Hootsuite social listening guide.

The bottom line

Sprout Social Listening is a capable, well-documented monitoring tool inside a suite you may already use. Its limits are mostly the networks' limits: hashtag-only public Instagram, Pages-only public Facebook, owned-page LinkedIn, owned-profile TikTok, no demographics and up to 30 days of backfill unless you pay for more. Because the price is unpublished and sits on top of a per-seat plan, often an annual contract that renews automatically, go in with a clear list of Topics and sources, get the quote in writing, and build your key Topics before the moment you need them.