You turned on Orbit last week. You picked three topics that felt right. Your first watchlist came back with 2,400 people and you have no idea what to do with it.
Or worse, you picked the wrong filters and the list looks like every cold database dump you have ever ignored.
This post is the playbook for building an Orbit audience you will actually work. Not a 20,000-row CSV. A tight, weekly-refreshed list of people who are doing something on the open web that matches the exact shape of a customer you want to close this quarter.
I am Elene on the Leadpipe team. I build these audiences with customers every week. The ones that convert are not the biggest. They are the ones with a narrow thesis, narrow filters, and a weekly cadence of action.
What an Orbit audience actually is
Leadpipe Orbit reads buying intent from a cross-site pixel network of 5 million websites and 60 billion signals. It tells you which people (not just companies) are researching a specific topic today, based on the pages they are reading across the open web, not just your site.
An Orbit audience is a saved query. You define it once. It refreshes daily. Every morning you get a fresh list of people who match your criteria.
The quality of that list depends entirely on how you define it. Get the topics and filters right and you get a weekly watchlist a single rep can work. Get them wrong and you get a list your team ignores.
The answer up front: narrow thesis, narrow filters, weekly cadence
Here is the shape of an Orbit audience that produces meetings:
| Component | Good answer | Bad answer |
|---|---|---|
| Topic count | 2 to 4 specific topics | 10+ broad topics |
| Topic specificity | ”HubSpot to Salesforce migration" | "CRM” |
| Company filters | Industry + size band + geography | No filters, “all B2B” |
| Role filters | 3 to 6 specific job-title patterns | ”Any decision maker” |
| Exclusions | Existing customers, competitors, partners | None |
| Intent score | 60+ for priority, 40+ for review | No threshold |
| Refresh cadence | Daily refresh, weekly review | Monthly manual pull |
| Audience size | 150 to 600 people per week | 5,000+ per week |
| Consumer | One named rep or campaign | ”The team” |
If any of those rows look like the right column for your current Orbit setup, this post is for you.
The playbook
Step 1: Write a one-sentence thesis before you pick a single topic
The thesis is the customer you want to close next quarter, in one sentence. Example:
Mid-market HubSpot customers who are actively researching migrating to Salesforce because their deal flow has outgrown HubSpot’s reporting.
That is a thesis. Now every topic and filter you pick has to support that specific shape. “Anyone researching CRM” is not a thesis. It is a category.
Three common theses that work:
- Replacement thesis. Companies using [Competitor X] researching alternatives. (e.g., Mailchimp users researching Klaviyo.)
- Category-entry thesis. Companies who have never bought your category, researching the category for the first time. (e.g., a fintech team researching “vendor risk management” for the first time.)
- Expansion thesis. Companies who bought adjacent tools and now need yours. (e.g., bought a data warehouse, now researching reverse ETL.)
Your thesis drives every decision downstream. Write it down. Keep it to one sentence.
Step 2: Pick 2 to 4 specific topics
Leadpipe’s intent API covers 20,000+ topics. That sounds like a lot of room to play, and it is exactly the room where most teams fail. They pick 8 broad topics and end up with a list that is equal parts “serious evaluator” and “someone who Googled the term once.”
The rule: topics should be narrower than your product category.
For the HubSpot-to-Salesforce migration thesis:
- Good: “Salesforce vs HubSpot,” “HubSpot CRM alternatives,” “HubSpot limitations,” “Salesforce migration”
- Bad: “CRM software,” “Sales software,” “Marketing automation”
See leadpipe-intent-api-20000-topics for how the topic taxonomy works.
Step 3: Filter firmographics to the 80% of your customer base
Filters are not optional. Without filters, intent scores you from the tail of the bell curve, and the tail is mostly irrelevant.
The filters that matter most:
| Filter | Typical setting |
|---|---|
| Industry | 3 to 5 industries that represent your top customer cohorts |
| Company size | An employee band (e.g., 50 to 500 FTE) or revenue band |
| Geography | Countries where you actually sell (US, CA, UK is typical) |
| Tech stack (if you have it) | Present or absent signals that match your thesis |
| Funding stage (optional) | For startups: seed, Series A, Series B, etc. |
Start tight. You can always relax later. A tight filter producing 150 high-fit people weekly is worth more than a loose filter producing 2,400 mixed-fit people.
Step 4: Pick 3 to 6 role patterns
Intent at the company level tells you the company is researching. Role filters tell you which person to reach.
For a sales-tool purchase, typical role patterns are:
- VP Sales / Chief Revenue Officer / Head of Sales
- Director of Revenue Operations / VP RevOps
- Sales Operations Manager
- Chief Marketing Officer (for bundled sales-marketing tools)
Six is the ceiling. More than six and you are diluting into people who cannot drive a purchase.
Step 5: Exclude existing customers, competitors, and partners
The single highest-ROI filter on any Orbit audience is the exclusion list. Without it, your sales team will email existing customers about why they should buy your product. This is the fastest way to lose internal credibility.
Upload three lists:
- Every customer domain (existing paying accounts)
- Every competitor domain
- Every partner or agency whose team shouldn’t get sales outreach
Leadpipe’s suppression / exclusion feature is one of its documented differentiators over alternatives like RB2B. Use it. See the exclusions post for more.
Step 6: Set an intent score threshold
Every person on the list has a 1 to 100 intent score. Treat it as a priority band, not a hard cutoff.
- 80+ : immediate outreach
- 60 to 79: priority queue for the rep
- 40 to 59: review, warm outbound sequence
- Below 40: rotate into nurture
Do not blast everyone above 40 with the same email. The message to an 85-intent person (“I saw your team is actively evaluating migration”) is different from the message to a 45-intent person (“Noticed your team is starting to explore CRM upgrades”). One is specific, one is exploratory.
Step 7: Set audience size expectations before launch
A well-built audience for a mid-market SaaS vendor in the US produces roughly 150 to 600 qualified people per week. If yours comes back with 5,000+, your filters are too loose. If it comes back with fewer than 30, your filters are too tight.
There is no prize for a big audience. The prize is a weekly watchlist that one rep can realistically act on.
Step 8: Assign an owner and a cadence
An audience without an owner is dead on arrival. Name one person. Give them a weekly review slot on their calendar (30 minutes every Monday works). That person:
- Pulls the weekly refresh
- Filters to the top 40 by intent score + role fit
- Hands 10 to 15 to the AE for personal outreach
- Routes the rest into a warm sequence
Weekly cadence is the sweet spot. Daily is too noisy for most teams. Monthly loses the freshness advantage of the daily refresh.
Sample audience definition (copy this)
Audience: "HubSpot customers exploring Salesforce migration"
Thesis: Mid-market HubSpot customers actively researching Salesforce
migration because deal flow has outgrown HubSpot.
Topics (intent):
• "Salesforce vs HubSpot"
• "HubSpot CRM alternatives"
• "HubSpot limitations"
• "Salesforce migration guide"
Firmographic filters:
• Industry: SaaS, Fintech, Professional Services
• Company size: 100 to 500 FTE
• Geography: US, CA, UK
• Tech stack (if available): HubSpot present
Role filters:
• VP Sales, CRO, Head of Sales
• VP RevOps, Director RevOps
• Sales Ops Manager
Exclusions:
• All current Leadpipe customer domains
• All Leadpipe competitor domains
• Leadpipe partner/agency domains
Intent score:
• 60+ for priority
• 40 to 59 for nurture
Refresh: Daily
Owner: @elene
Review: Monday 10:00am (30 min)
Output: Top 15 → AE for personal outreach
Rest → warm sequence (3-email touch)Common failure modes
Too broad. Picking “CRM” as a topic. You will drown in generic researchers who are not near a buy.
Too many topics. Picking 12 topics and “OR’ing” them together. You end up with a surface audience where no one is deeply in the funnel.
No exclusions. Emailing existing customers about why they should buy your product. Uninstall your credibility in a week.
No owner. The audience refreshes daily and nobody looks at it. The value is zero.
Uniform outreach. Treating an 85-intent person the same as a 45-intent person with the same template. The 85 converts; the 45 unsubscribes.
Ignoring EU/UK. Leadpipe defaults to company-level data for EU/UK visitors. If your audience includes EU companies, your outreach at the person level requires more care. Send a company-addressed message first, or confirm consent before sending person-level email.
Conflating Orbit with visitor identification. Orbit is intent across the open web. Visitor identification is about your own site. They are complementary, not the same. See intent data vs visitor identification.
Measurement: how to know it is working
Three leading indicators after 30 days:
- Reply rate from first-touch emails above 10%. Below that, your topics or your message are off.
- Meetings booked per week per audience at 3 to 8. This is the output benchmark for a well-sized audience.
- SDR says the list is workable. Qualitatively, your rep should not be complaining that the list is junk.
Two lagging indicators after 60 to 90 days:
- Opportunities from Orbit-sourced contacts. At least 2 to 5 per month per audience.
- Close rate from Orbit-sourced opps at or above cold-outbound. Orbit is outbound-warm, not outbound-cold. Close rate should beat cold by 1.5 to 2x.
One trap. Do not let Orbit replace your inbound. It runs in parallel. The intent + visitor-ID combo is the full-stack approach. Orbit finds them off-site, identification catches them when they show up on-site.
Where Orbit sits in the rest of your stack
Orbit pairs with:
- Visitor identification for on-site activity
- Slack alerts for real-time routing
- HubSpot or Salesforce for pipeline capture
- Warm outbound sequences for the longer tail
The failure mode is treating any of these tools in isolation. Orbit without visitor ID misses the signal when the prospect shows up on your site. Visitor ID without Orbit misses the signal when they are researching but haven’t landed yet. Together they produce a complete picture.
See also orbit audiences every saas company should build for audience ideas by industry, and person-level intent data how it works for the mechanics.
Build your first Orbit audience in under 20 minutes. 5M-site pixel network, 60B+ intent signals, daily refresh, person-level resolution. Start free →
