The whole point of intent data is catching a buyer before they pick a shortlist. If your tool surfaces accounts that already booked demos with three of your competitors, it is not giving you intent data. It is giving you a lost deal report.
I am George, founder of Leadpipe. Teams evaluating Orbit ask the same question in slightly different words: how early does this actually show me a buyer? Not “when does a dashboard light up a week after an RFP.” Early, as in the first few sessions on the topic, before the buyer knows they are a buyer. This post is the honest answer.
The short answer
Orbit typically surfaces a person 2 to 5 days after they begin meaningful research on a category, and sometimes inside the first session if the signal is strong enough. “Meaningful” means multiple pages, more than one site, and topic-specific content (not a single incidental pageview). The earliest dependable detection point is usually the second session, because the first session alone is hard to separate from noise without a second data point. Detection speed depends on how much of the buyer’s research happens on sites where our pixel network runs, how distinctive the topic is, and where you set your score threshold.
That is the one-paragraph version. The rest of this post is the timeline, the mechanics, and the honest limits.
The buyer journey, mapped to Orbit detection
Here is the stylized path a B2B buyer takes when entering a new category, and when Orbit typically sees them.
| Phase | Buyer behavior | Earliest Orbit visibility | Typical score range |
|---|---|---|---|
| Phase 0: problem awareness | Reads a blog post about their pain | Usually invisible | under 50 |
| Phase 1: category discovery | Searches for “what is X”, reads 2-3 articles, maybe a vendor homepage | Day 1-3 of activity | 50-65 |
| Phase 2: vendor list | Visits 3-5 vendor sites, starts reading comparison posts | Day 2-5 of activity | 65-79 |
| Phase 3: shortlist | Pricing pages, “X vs Y” posts, feature deep dives | Day 4-10 of activity | 80-92 |
| Phase 4: late-stage | Repeat pricing visits, trial page, demo request pages | Day 7-21 | 90-100 |
Most teams set their Orbit audience floor at 70. That puts you right at the start of Phase 2: vendor list. That is the ideal intercept window. Early enough to influence the shortlist, late enough that the person is actually a buyer.
Raise the floor to 80 and you are reading Phase 3. Drop it to 60 and you are catching Phase 1. The threshold is the knob.
What makes “early” possible
Early detection requires three things to line up.
1. The signals need to reach us. Our cross-site pixel network runs on roughly 5 million sites and collects about 60 billion signals per day. That footprint is wide enough that most B2B research behavior passes through at least some of it. Vendor sites, comparison sites, review sites, long-tail publications, and content hubs are all represented. But we do not see every page on the internet. If a buyer spends their entire research cycle on Reddit, Google search result pages, and a closed Slack community, we will see less.
2. The topic has to be well-defined. Generic topics like “software” are useless for early detection because every page triggers them. Specific topics like “Salesforce alternatives” or “HIPAA-compliant file sharing” produce cleaner signals. The 20,810-topic taxonomy in Orbit is built around specificity, which is why topic selection matters more than anything else in setup.
3. The scoring has to punish noise. A single pageview is not intent. Our score blends page count, dwell time, return visits, cross-site diversity, and recency. A score of 80 requires more than one visit, more than one page, and usually more than one site.
When those three things align, you see a person early. When they do not (niche category, private communities, single long-read article) you see them late or not at all. We are not pretending otherwise.
A worked example: catching a CRM buyer on day 3
Here is what a real Phase 2 detection looks like. This is a reconstructed example from a pattern we see often.
Day 1. A VP of Marketing reads a blog post on a vendor site titled “when to move off HubSpot.” Single pageview, 4 minutes dwell time. Our classifier tags the page as “CRM Software” and “HubSpot Alternatives.” The score for this person on those topics is around 44. Below threshold. Invisible in most Orbit audiences.
Day 2. Same person visits two comparison sites, reads “HubSpot vs Salesforce” and “best CRMs for mid-market.” About 12 minutes of total dwell across 4 pages. Cross-site signal appears. Score rises to 68 on “CRM Software,” 72 on “HubSpot Alternatives.” This is a visitor with at least one audience if your threshold is 70.
Day 3. Vendor pricing page on a third CRM site. Short visit but the page type is high weight. Score jumps to 81. If you run an audience for “CRM Software” with a floor of 80 and filter to seniority VP+, this person surfaces in your next daily run.
Total elapsed time from the first on-topic pageview to a strong Orbit signal: about 48 hours.
Where the limits actually are
Being honest about what we do not catch matters more than hype.
- Pure Google search behavior without clickthrough to a site in our network produces no signal. If someone types “best CRM” into Google and closes the tab, we do not see it.
- Private communities and paid content are mostly invisible. Reddit, Slack, Discord, gated analyst reports in walled portals.
- Single-article researchers who read one piece and never return are visible at low scores but usually drop off before crossing the threshold.
- Freshly minted topics (a competitor that launched last month) take a few days to accumulate signal density before early detection stabilizes.
- EU/UK users default to company-level resolution only. Person-level requires affirmative consent.
We do not score a person who has had zero interaction with the pixel network. We do not fabricate signals to fill in gaps. If the answer is “we do not see this person yet,” Orbit says so by not surfacing them.
How to tune for early detection
If early catch is the priority, here is a practical configuration.
- Pick narrow topics. Competitor-named topics, specific product categories, “alternatives to X.” Avoid top-of-funnel generic topics.
- Use topic overlap. Set
minTopicOverlap: 2across two or three related topics. A person showing up on three adjacent topics is a stronger early signal than one topic alone. - Lower the score threshold to 65. You will get more noise, but more Phase 1 and Phase 2 catches. Pair this with a tighter ICP filter so the volume stays manageable.
- Run the audience daily. Orbit refreshes audiences every 24 hours. New signals appearing today show up in tomorrow’s run. If you pull weekly, you lose the early window.
- Cross-reference with visitor identification. If you also run Leadpipe visitor ID on your site, intent signals from Orbit plus a first visit to your pricing page is a very high-confidence intercept point.
Comparison: Orbit vs publisher co-op vs ABM platforms
Detection speed varies by architecture, not by marketing claim.
| Source | Detection granularity | Earliest visibility for a new buyer |
|---|---|---|
| Bombora (publisher co-op) | Company surge | 1-2 weeks of sustained company-level activity |
| 6sense / Demandbase | Blended signals, company-level | 1-3 weeks, often late in the shortlist phase |
| G2 intent | Review site behavior, often late | Usually Phase 3 or later |
| Clearbit Reveal (HubSpot Breeze) | Your own site traffic, company-level | Only after the buyer visits your site |
| LinkedIn lookalikes | Not intent, modeled similarity | No detection, just audience expansion |
| Orbit | Person-level, keyword topic, daily refresh | 2-5 days into meaningful research |
Bombora is legitimately good at catching high-volume category surges in established markets. 6sense has strong account-level orchestration. LinkedIn lookalikes are helpful for reach, not for intent. Each serves a different purpose.
Orbit’s specific advantage is the combination of person-level resolution, keyword-level specificity, and daily refresh. That combination is what allows Phase 2 detection. For the larger framing, see Orbit vs Bombora and the person-level intent data primer.
The playbook: act on early signals
Catching a buyer in Phase 2 is only useful if your team reacts fast. A 72-hour-old signal is gold. A three-week-old signal is a lost deal. Some rules we use with customers:
- Daily alerting. Orbit audiences refresh every 24 hours. Wire a webhook or a daily CSV export into Slack or your SDR queue. Do not batch weekly.
- Personalization by topic. A person surfacing on “CRM Migration” gets a different opener than someone on “CRM Software” generally. The topic is in the payload.
- Respect the intent window. A buyer with a score of 78 today might spike to 92 next week or decay to 50. Reach out now, do not wait for “perfect” scores.
- Tier the response by seniority. VP and C-Suite get personalized founder-level outreach. Director and Manager go to standard SDR sequences.
When detection is slow
If your Orbit audience is coming in later than you expected, four things usually explain it.
- Topics too broad. Narrow them.
- Score threshold too high. Drop it to 70 and add a stricter ICP filter instead.
- ICP filter too tight. Loosen department or seniority by one bucket.
- Category so niche the pixel footprint is thin. This is rare but real. In those cases, Orbit is a supporting signal, not a primary one.
The honest bottom line
Orbit catches most B2B buyers 2 to 5 days into their research cycle, which is usually before they contact a vendor. It will not catch everyone, especially buyers whose research happens in private channels. It will not invent signals for people who are not browsing. Inside those limits, daily refresh plus person-level resolution plus keyword-specific topics is the earliest honest detection point I have seen in the category.
If you want the full taxonomy mechanics, read what’s inside an Orbit intent topic and the buyer intent glossary entry. If you want the mechanics of the 1-100 scoring layer, read reading the Orbit intent score.
Orbit resolves intent at the person level against a deterministic identity graph — the difference between “an account is in-market” and “this director at this account is researching today.” Try Orbit →
