TL;DR
The most reliable way to detect that a company is trialling a competitor is to observe the specific technical signatures that competitor's product leaves on the trialist's infrastructure: DNS records, JavaScript loaded on their site, subdomain patterns, integration endpoints, and support-widget IDs. Behavioural proxies (job postings, LinkedIn mentions, review-site visits) are noisier and slower. The four detection methods below are ranked by fidelity, with the important constraint that only the top two produce signals fresh enough to act on inside the 14-to-30-day evaluation window when the outreach conversation is still open.
Why competitor trials are the highest-signal event in B2B SaaS
A competitor trial is the single strongest signal in outbound because it collapses three uncertainties at once: is this account in the market, is our category actually a fit, and are they willing to change vendors? A live competitor trial answers all three affirmatively.
Every other buying signal answers only one or two of those questions. A pricing-page visit tells you the account has some curiosity, not that they are actually evaluating. A content download tells you the account is topic-aware, not that they are close to a decision. A conference badge scan tells you they attended, not that they are shopping.
A trial start tells you the account is doing the work of switching. The evaluation is underway, the internal alignment has cleared enough hurdles to authorise the trial, and the buyer is actively inside a 14 to 30 day window where their attention is on your category. Outreach inside that window converts at multiples of the same outreach outside it. See Timing Intelligence 101 for why the temporal layer matters and what depends on getting it right.
What a competitor trial actually looks like (technically)
When a SaaS company starts trialling a new tool, the tool leaves technical signatures on their public infrastructure. The signatures are not private data; they are the observable side effects of installation, and they are inspectable without any access to the trialist's internal systems.
JavaScript loaded on the trialist's site. Most B2B SaaS tools inject a JavaScript SDK on the customer's website (analytics, live chat, product tour, feedback, support widget). The SDK is loaded from the vendor's domain and often includes a customer-specific ID in the request path. Watching the JS load pattern is the highest-fidelity signal for tools that live in the customer's frontend.
DNS records and subdomain patterns. Many SaaS tools require the customer to create a subdomain (support.customer.com, chat.customer.com, academy.customer.com) that CNAMEs to the vendor's infrastructure. The DNS record itself is public and durable, and adding one is a strong indicator of installation, not just evaluation.
Integration endpoints. Tools that live in the backend (CRM connectors, sync tools, workflow platforms) often expose OAuth callback URLs, webhook endpoints, or public API integration paths that name the vendor. Slower to detect than frontend signatures, but often more decisive when they appear.
Support-widget IDs and customer-specific paths. Live chat widgets, help centres, and knowledge bases hosted by the vendor use customer-specific IDs that appear in the widget's initialisation code. These IDs are the cleanest possible confirmation of an active customer, not just a trialist, because they persist as long as the tool is running.
The critical distinction: signatures that appear on day one of a trial versus signatures that only appear once installation is complete. Fast detection means catching a trial in the first 48 hours. Slow detection means catching an already-installed tool three weeks later. Both are useful, but only fast detection lands you inside the 14-to-30-day evaluation window.
The four detection methods, ranked by fidelity
Four ways to detect a competitor trial, from strongest to weakest. Fidelity here means the ratio of true positives to false positives, and the latency between "trial actually started" and "signal fires."
Method 1: Technical signature monitoring. Continuously scan target-account domains for the technical signatures listed above and fire an event when one appears that was not present in the previous scan. This is the highest-fidelity method and the only one that reliably catches trials on day one. Fidelity depends on the coverage set (does the vendor being trialled have a monitored signature?) and the scan cadence (daily is table stakes; weekly is too slow to matter). This is the method underneath most modern Subscription Intelligence tools; see What Is Subscription Intelligence? for the category.
Method 2: Review-site activity monitoring. Watching G2, Capterra, TrustRadius, and similar sites for a target account's employees writing reviews of a specific vendor. High confidence when it fires (reviews correlate strongly with actual usage), low coverage (most companies do not write reviews), and inconsistent latency (reviews often lag installation by weeks). Useful as a corroborating signal, weak as a primary trigger.
Method 3: Job-posting inference. A target account posts a role that explicitly names a competitor tool ("must have hands-on experience with Zendesk"). Reasonable proxy for tool consideration or use, but ambiguous: does it mean the tool is already installed, or that the account is considering it? Latency is long (roles are open for weeks before they signal a real decision). Useful for account-level qualification, weak for timing.
Method 4: LinkedIn / social mention scanning. Watching for employees at a target account posting or engaging with a competitor's content. Directionally useful, but heavily noise-prone. A rep at the account "liking" a competitor's LinkedIn post is not evidence of a trial. Best treated as a nurture-side input, not a prioritisation input.
The practical stack for a serious detection motion combines Method 1 as the primary trigger with Method 2 as a corroborator. Methods 3 and 4 belong in the nurture layer.
The 14-to-30-day window: why timing decides the outcome
Once a trial fires, the outreach window is short. Most SaaS trials run 14 or 30 days. Inside that window, the buyer is actively comparing, budget conversations are in motion, and the vendor being trialled has a rapidly-closing lead.
Outreach inside the window has three properties that outreach outside the window does not. The buyer is expecting to be sold to. The buyer's attention is on your category. The buyer's alternative (staying with their current vendor) is being actively questioned by the trial itself. All three of those tail off the moment the trial ends and a decision is made.
Miss the window and the account goes cold for the length of whatever contract the winning vendor signs, usually a year or three. That is the specific cost of a slow detection method: not "we missed one deal," but "we cannot enter this account's evaluation again for 12 to 36 months."
The math on that is why detection latency matters so much. A daily-refresh signal that fires on day two of a 30-day trial gives you 28 days to act. A weekly-refresh signal that fires on day nine gives you 21 days. A monthly signal that fires on day 25 gives you five. Same underlying data, wildly different conversion outcomes.
What to do when a trial detection fires
Four moves. In this order.
1. Verify the signal. Every detection method has a false positive rate. Before a rep touches the account, look at the underlying evidence: what technical signature fired, when, and how confident is the detection. A well-designed Subscription Intelligence tool exposes this directly. See Explainable AI in B2B Sales for why inspectable evidence beats explanation-shaped text on this specific problem.
2. Frame the outreach around the specific competitor being trialled. Generic "we heard you might be evaluating" outreach lands as spam. Specific "we noticed you started a trial of X, here is where our category tends to deliver better outcomes on Y" outreach reads as informed. The difference is what the rep does in the two seconds before hitting send.
3. Prioritise on the trial plus the ICP fit, not the trial alone. A trial detection on a company that is not in your ICP is a distraction, not an opportunity. Combine the trial signal with your existing ICP scoring and the Win Probability layer that scores how winnable an account historically is against that specific competitor. See Qualified Opportunity: the three-part definition for the full ICP-fit + purchase-intent + win-probability framework.
4. Log the outcome, whatever it is. The detection layer is a system; systems improve when their outputs are measured. If a trial detection produced a meeting, note it. If it did not, note that too. Six months of outcome tags is how a team distinguishes signals worth acting on from signals worth deprioritising.
Frequently asked questions
How do you detect that a company is trialling a competitor?
The most reliable method is monitoring technical signatures the competitor's product leaves on the trialist's public infrastructure - JavaScript loaded on their site, DNS records or subdomains that CNAME to the vendor, OAuth callback URLs, and customer-specific widget IDs. When one of these signatures appears where it was not present in the previous scan, a trial has started. Behavioural proxies (job postings, review-site activity, LinkedIn mentions) are noisier and typically lag installation by days to weeks.
How fast can a competitor trial be detected?
Technical signature monitoring with a daily scan cadence catches most trials within 24 to 48 hours of installation. Weekly scans catch them within about a week. Monthly scans routinely miss the 14-to-30-day trial window entirely. Freshness is the whole point: a signal that fires after the trial has ended is a retrospective, not a buying signal.
What is the difference between detecting a trial and detecting an installed customer?
Some technical signatures (initial DNS records, first JavaScript load) appear on day one of a trial. Others (customer-specific IDs, sustained integration endpoints) only appear once the tool is fully installed and configured. Trial detection requires the first class of signatures; installed-customer detection uses both. Only trial detection lands you inside the buying window; installed-customer detection is more useful for renewal-timing plays.
Can I detect competitor trials at scale without accessing the trialist's data?
Yes. Every method described here relies exclusively on public signatures. Nothing scans private data, nothing accesses the trialist's internal systems, and nothing depends on a partnership with the vendor being trialled. The signatures are observable side effects of installation, not private records.
What is the false positive rate on competitor trial detection?
Rate depends on the detection method and the vendor's coverage set. Method 1 (technical signature monitoring) on well-covered vendors runs at 90 to 95% precision in MarketSizer's own dataset. Method 2 (review-site) runs higher precision but much lower recall. Any vendor claiming 100% accuracy on trial detection is either lying or does not understand their own product; the honest number is high nineties and the honest thing to do is expose the confidence level per signal so reps can act accordingly.
What tool detects competitor trials for B2B GTM teams?
Subscription Intelligence platforms detect competitor trials as one of their primary signal types. MarketSizer covers 140+ Customer Support, Live Chat, and adjacent Martech products with daily-refresh technical signature monitoring; the Bloomberry and TheirStack products cover overlapping but different vendor sets. Choose based on which vendors matter most in your competitive set and what the tool's coverage looks like against your target account list, not on marketing claims.
Recommended reading
- What Is Subscription Intelligence? - the underlying data category that makes trial detection possible at scale.
- Timing Intelligence 101 - the pillar on why buying windows decide outreach outcomes.
- Qualified Opportunity: the three-part definition - the ICP-fit + intent + win-probability framework that trial detection feeds into.
- Explainable AI in B2B Sales - why inspectable evidence matters for verifying a trial signal before acting.
- The Signal-Led GTM Playbook - the operating model that runs trial detection as one of its four core moves.