Signals — Spot What's Working Before Everyone Else
Get alerted when an account you follow has a post doing far better than usual, or when new content matches something you told us to watch for.
Discover shows you everything the accounts and websites you follow are posting. Signals pick out the few things worth your attention. You don’t have to scroll through it all yourself.
There are two kinds of signals:
- Breakout posts: a post from an account you follow is doing much better than that account usually does. We find these for you automatically.
- Custom rule matches: new content matches something you told us to watch for, in your own words. For example, “news about changes to Instagram’s algorithm”.
You’ll find them in Discover, on the Signals tab. Click Signal in the sidebar to go straight there.
Signals come from your Discover sources. To choose which accounts, websites and Reddit communities to watch, see Discover Sources.
Quick start (5 minutes)
- Open your brand’s Discover sources.
- Under Social Media, add 3 to 5 accounts you want to learn from. Pick competitors, brands you admire, or leaders in your industry.
- Under Signal Rules, click Add.
- Type a Name, like “Competitor price changes”.
- Type the Criteria, like “A competitor announcing new prices or a new plan. Not general sales or discount codes.”
- Under Watch these sources, tick the sources the rule should read.
- Click Add Rule.
- Check back tomorrow. New signals show up once a day.
Breakout posts
Every account has its own normal. 500 likes is amazing for a local bakery. For a national brand, it’s a quiet day. So we don’t compare raw numbers. We compare each post to that account’s own usual engagement.
How we spot a breakout
- We work out each account’s usual engagement rate from its posts over the last 60 days. Engagement counts likes, comments, shares, saves and similar actions, compared with the account’s number of followers.
- We compare each recent post against that usual rate.
- If a post does at least 1.5 times better than usual, it becomes a signal.
Example: A café’s posts usually get engagement equal to 2% of its followers. A new Reel gets 8%. That’s 4 times its usual, so it shows as Breakout · 4.0x.
How much better is shown as a number, like 4.2x, and a label:
| Label | How much better than usual | What it usually means |
|---|---|---|
| Emerging | 1.5x to 3x | Doing well. Worth a look. |
| Breakout | 3x to 5x | Clearly working. Study it. |
| Strong | 5x to 8x | A standout for this account. |
| Viral | 8x or more | Rare. Something really clicked. |
Breakout posts are re-checked every day as they get more engagement. A post can move up from Emerging to Viral over a few days.
Trending topics
Sometimes breakout posts from different accounts cover the same topic at about the same time. The card then shows a purple badge, like “4 accounts · 3.1x”, instead of the label. The account count only includes accounts your brand follows. Hover over the badge to see the topic. A trending topic often means the whole industry is talking about something. That’s a good moment to join in.
Low confidence
Some breakout posts show a Low confidence badge. This means the account posted fewer than 20 times in the last 60 days. Its “usual” is based on only a few posts, so it’s less reliable. Hover over the badge to see how many posts it’s based on.
Which accounts are included
- Accounts you’ve added under Social Media in your Discover sources.
- Instagram, TikTok, YouTube, Facebook and LinkedIn.
- Accounts that post regularly: at least 10 posts in the last 30 days. For accounts that post less often, at least 10 posts in the last 60 days.
Your own channels aren’t shown in breakout posts. Breakout posts are about learning from others.
Custom rules
Breakout posts tell you what’s performing. Custom rules tell you when something relevant appears, whether or not it’s popular yet.
You describe what you’re looking for in plain words. Every day, Orsa AI reads the new content from the sources you picked. It decides whether each item really matches. There are no keywords to set up. Orsa AI understands what you mean. A rule for “a competitor announcing a price change” also catches “we’re making our plans more affordable”.
Create a rule
- Open your brand’s Discover sources.
- Scroll to Signal Rules and click Add.
- Type a Name you’ll recognize, like “Algorithm news”. The name is shown on every match.
- Type the Criteria: what to watch for, in a sentence or two. For example: “News about changes to how Instagram or TikTok decide which posts to show. Not general tips.” (See Writing good criteria.)
- Under Watch these sources, tick the sources this rule should read. You can pick Websites, Web Watchlists, Social Accounts, Subreddits and Reddit Queries.
- If you ticked social accounts, choose an Engagement level:
- Any: check every new post.
- Moderate: only posts that are already a breakout post (Emerging or higher).
- High: only posts doing at least 3 times better than usual.
- Leave Active ticked.
- Click Add Rule.
Example: A dog groomer follows three local pet shops on Instagram. They create a rule named “Grooming offers” with the criteria “A pet shop announcing a grooming discount, a new grooming service or new grooming prices.” They tick the three accounts and leave the engagement level on Any.
Manage your rules
- Use the switch on a rule to pause it or turn it back on. A paused rule shows Paused and doesn’t check anything.
- Hover over a rule and click the pencil to edit it.
- Hover over a rule and click the bin to delete it. You’ll be asked to confirm. Deleting a rule also removes all of its matches.
- A rule that says “No sources selected — this rule will never run” needs at least one source ticked.
Writing good criteria
Orsa AI is told to be strict. Only clear matches show up in Signals. Clear, specific criteria get you better matches.
- Say what you want, not just a topic. “A competitor launching a new product or service” works better than “competitor products”.
- Say what to leave out. “Price changes from competitors. Not general sales or discount codes.”
- One idea per rule. Want to watch two things? Make two rules. Each match then tells you clearly why it’s there.
- Use your own words. No need for keywords, quotes or special symbols.
Which content gets checked
- Websites: new blog posts and articles.
- Web watchlists: new mentions found by your watchlists.
- Social accounts: new posts, once Orsa AI has finished reading them. Photo posts are usually ready within a day of being collected. Posts Orsa AI couldn’t read after two days, like many videos, are checked using their caption.
- Subreddits and Reddit queries: new Reddit posts. Reddit is collected once a week, so Reddit matches arrive weekly rather than daily.
Each rule checks content collected from 10 days before you created it onwards. A new rule starts with a little recent history instead of an empty list.
Using the Signals tab
Open Discover and click Signals. Breakout posts and rule matches are shown together, newest first. Click Load More at the bottom to see older ones.
On each card
- A colored badge. It shows the breakout label (like Strong · 5.4x), a trending topic, or the name of the rule that matched in blue. Hover over a rule badge to see why Orsa AI thinks it matches.
- Posted shows when the content was published.
- Detected shows when it became a signal.
- An Unread dot until you open it.
What you can do
- Click a card to open the post details on the side. This also marks it as read. See Discover Feed.
- Click the arrow icon to open the original post or page.
- Click the star to save it to your Swipe File. You can add a note on why it works, like “Strong hook in the first second”. See Swipe File.
- Click the lightbulb to create a content idea from it. The idea links back to this post as inspiration. The card then shows the ideas you made from it.
- Click Mark all as read to clear everything that matches your current filters.
Read and unread are tracked for each person. Marking a signal as read doesn’t mark it for your teammates.
Filtering
- The All Signals menu switches between everything, Breakout Posts only, or Custom Rules only.
- Unread shows only what you haven’t opened yet.
- Filters narrows by source type, account, channel or website, and by two
dates:
- Posted: when the content was published.
- Detected: when it became a signal. Use Detected: Last 7 Days for “what’s new this week”.
Your filters are remembered for each brand in this browser, so they’re still set next time you visit. Click Clear all to reset them.
Notifications
When new breakout posts are found, you get one notification in the app, like “5 new signals detected”. Clicking it opens the Signals tab. If you’re in more than one brand, it’s a single notification covering all of them.
New custom rule matches send their own notification, like “3 new rule matches found”. These come once a day when there are any. Social posts are counted in one notification. Websites, watchlists and Reddit are counted in another. So on some days you may get two.
To turn signal notifications off:
- Go to your profile.
- Open Notifications.
- Switch off Signals.
Ideas for using signals
Learn from competitors Add your 3 to 5 closest competitors. When one of them has a Breakout or Strong post, study it: the hook, the format, the topic, the timing. Save it to your Swipe File with a note. Then use the lightbulb to turn it into an idea of your own.
Example: A café follows four other local cafés. One has a Strong · 6.1x Reel showing a barista pouring latte art in slow motion. The café saves it with the note “Slow-mo process shots: try with our pastry lamination”. Then it creates an idea for next week.
Jump on trends early Look for the purple trending badge. When several accounts break out on the same topic in the same week, there’s an opening. Post your take while people still care.
Example: A fitness coach sees “5 accounts · 3.4x” on posts about “walking pads for desk workers”. They make a carousel on 3 desk-friendly stretches the same week.
Stay on top of industry news Add the industry blogs or news sites you’d normally check by hand. Then create a rule for the news that affects you.
Example: A social media agency adds a few marketing news sites and creates: Name: Algorithm news. Criteria: “News about changes to how Instagram, TikTok or LinkedIn decide which posts to show people. Not general tips or how-to articles.”
Watch for mentions and opportunities Point a rule at your web watchlists. It catches people talking about your brand, your product category or a problem you solve.
Example: A wedding photographer creates: Name: Venue openings. Criteria: “A new wedding venue opening, or an existing venue reopening after a renovation, in or near Austin.”
Hear what people ask on Reddit Point a rule at a subreddit or Reddit search where your customers hang out. You’ll hear about questions and complaints you could answer with your own content.
Example: A bike shop watches r/bikewrench. Its rule: Name: Repair questions. Criteria: “Someone asking how to fix or maintain their bike, or where to get it repaired.” Each match is a ready-made topic for a how-to post.
Find out which of their posts really perform Combine both kinds of signal. Create a rule that watches competitor accounts with the engagement level set to High. You only hear about their posts on a specific topic that also performed.
Example: A skincare brand creates: Name: Competitor launches that landed. Criteria: “A post announcing or showing a new product.” Engagement level: High. The result is a short list of launch posts that clearly worked, ready to learn from.
Build a weekly routine Once a week, open the Signals tab and:
- Set Detected: Last 7 Days and Unread.
- Save the best two or three to your Swipe File.
- Turn one into an idea.
- Click Mark all as read.
Tips for the best results
Follow accounts that post often. Breakout posts only work for accounts with at least 10 posts in the last 30 to 60 days. The account also needs a public follower count. An account that posts twice a month will never show breakout posts. Pick active accounts, like a competitor that posts most days.
Give each account time to build its “usual”. Accounts with fewer than 20 posts in the last 60 days show Low confidence. Use those signals as a hint, not a verdict. Look at the account’s other posts before you copy what it did.
Fill in your brand details. When Orsa AI checks content against a rule, it also reads your brand’s About section, pillars and audiences. Clear brand details help Orsa AI judge what’s relevant to you. For example, a rule for “local events we could take part in” works better when your About section says which city you’re in.
Be specific in your criteria. Only clear matches show up in Signals. Matches Orsa AI isn’t sure about are left out. Vague criteria like “marketing” rarely produce a clear match. “A competitor launching a loyalty program” does.
Use the engagement level to cut the noise. A rule set to Any on busy social accounts can match a lot of posts. Set it to Moderate or High to only hear about posts that are already performing.
Make a new rule instead of rewriting an old one. Each piece of content is checked once per rule. Changing a rule’s criteria only affects new content. To re-check the last 10 days against new wording, create a new rule.
Pause a rule instead of deleting it. Deleting a rule removes all of its matches. Pausing keeps them. When you turn a paused rule back on, it also checks the content collected while it was paused.
Save the good ones before they age out. Breakout posts drop off the Signals tab once the post is more than a month or two old. Save the ones you want to keep to your Swipe File.
Keep your Discover sources tidy. If you remove an account from your Discover sources, its breakout posts leave your Signals tab too. Remove accounts you no longer care about, and your Signals stay focused.
Common questions
I don’t see any signals yet. Signals are checked once a day, so give it a day after adding sources. Breakout posts also need accounts that post regularly. An account that posts once a month won’t have enough history. To see when each kind of source updates, see How Discover Updates.
What does “Also check comments on these accounts’ posts” do? Nothing yet. Discover doesn’t collect comments from social accounts at the moment, so this option doesn’t add any matches. Rules still check the posts themselves.
An account I follow never has breakout posts.
- Check that it posts at least 10 times every month or two.
- Check that it has a public follower count. Without followers, we can’t work out its engagement.
- It may simply not have had a standout post recently.
My rule isn’t matching anything.
- Check the rule is switched on and has at least one source ticked.
- Make the criteria a little broader. Orsa AI is strict on purpose, and only clear matches are shown.
- If you set the engagement level to Moderate or High, only posts that are already performing are checked. Try Any.
- New matches show up about a day after the content is collected.
I edited my rule, but old content didn’t get re-checked. Each piece of content is checked once per rule. Changes apply to new content from then on. To check recent content against new criteria, create a new rule instead.
I ticked subreddits or Reddit queries in a rule, but nothing matches from Reddit yet. Reddit posts are collected once a week, so give it up to a week. Rules only check Reddit posts collected from 10 days before the rule was created onwards.
Can I see breakout posts from my own accounts? Not on the Signals tab. It only shows accounts you follow in Discover.
Where did an older breakout post go? Breakout posts leave the list once the post is more than a month or two old. Save the ones you want to keep to your Swipe File.
A teammate says they’ve read everything, but I still see unread signals. Read and unread are tracked for each person. Each teammate marks signals as read for themselves.
Is there a limit?
- There’s no limit on the number of rules.
- Each source you add counts towards your plan’s Discover source limit. That includes accounts, websites, watchlists, subreddits and Reddit queries.
- Each day, a rule checks up to 200 new items from each kind of source. On a very busy day, the rest are checked the next day.
Was this helpful? Tell us what worked or what was missing athello@orsa.social.