The Stratridge competitor-signal archive holds 10,400 tracked competitor shifts across 214 B2B SaaS companies, collected from Q1 2023 through Q4 2025. The archive was assembled initially for operational purposes — to help our own monitoring clients benchmark their programs — but the aggregate produced patterns we didn't expect. The three findings below are the ones that have changed how we recommend clients build monitoring programs, and the ones most worth knowing if you operate a monitoring program of your own.
The aggregate distribution
Before the patterns, the distribution of signal types. The 10,400 shifts, by type:
The distribution confirms the intuition that feature launches dominate the tracked-signal volume. They account for 30% of signals. Pricing adjustments and homepage copy changes together account for another 31%. The remaining 39% is spread across eight other categories.
What the distribution does not show — and what the follow-up analysis did — is that signal volume and signal predictive value are very different things.
Finding 1 · Feature launches are mostly noise
Of the 3,120 feature launches tracked, 1,840 had 12-month follow-up data (win-rate change, pipeline impact, or analyst coverage shift). Of those 1,840, exactly 142 (7.7%) produced a measurable market outcome in favor of the launching competitor within 12 months. The other 92% either produced no measurable shift or produced a shift within normal market variance.
The implication: the monitoring programs that weight feature launches as primary signals are chasing noise 92% of the time. A program that ignored every feature launch and acted only on other signal types would have missed 7.7% of legitimate threats — but freed up the PMM hours that 92% of feature-launch responses consumed.
We halved our feature-launch response volume after seeing this data. We stopped updating battle cards for every competitor feature and started updating them only when the feature was paired with a pricing change or customer advocacy. Our battle cards got fresher — because we were updating fewer of them, more carefully.
The exceptions — the 7.7% of feature launches that did produce outcomes — had specific properties. They were almost always paired with either (a) a pricing change within 60 days, (b) named customer advocacy within 90 days, or (c) analyst coverage within 6 months. Isolated feature launches — a product update with no surrounding signal — rarely mattered. The pattern is the signal, not the feature.
Finding 2 · Careers-page patterns are the highest-return early signal
The most striking finding in the archive is the predictive value of careers-page patterns. Of the 1,180 tracked careers-page shifts, the subset that followed specific patterns (three or more aligned hires within a 90-day window) produced a measurable market-outcome prediction at a 73% hit rate within 9 months.
Specifically: when a competitor hired three or more roles aligned with a specific market direction (enterprise sales, vertical specialization, AI/ML infrastructure), the competitor launched a material initiative in that direction within 9 months in 73% of cases. The remaining 27% either pivoted away from the direction or the hires became standalone operational additions without public launch.
For most B2B SaaS monitoring programs, careers-page monitoring is a tier-two activity. The archive data suggests it should be tier-one. The lead time advantage — 9 months before public announcement — is uniquely valuable; no other signal type in the archive produces comparable lead time at comparable accuracy.
The operational implication: monitoring programs should reweight their time allocation. The PMM hours spent on feature-launch triage should be redirected to careers-page pattern tracking. The volume-to-value ratio is substantially better.
Finding 3 · Three patterns show up in every successful competitive response
Of the 214 companies in the archive, 89 experienced a competitive threat substantial enough to require a coordinated response during the tracking window. Of those 89, 34 successfully neutralized the threat (measured by sustained or improved win rate 12 months post-response), and 55 did not.
The archive compared the response patterns of the 34 successful and 55 unsuccessful responses. Three patterns appeared in nearly every successful response and rarely in the unsuccessful ones.
Pattern 1 · Response time was not fast, but decision time was
The successful responses took a median of 6–12 weeks from signal-detection to public response. The unsuccessful responses broke into two groups — those that responded in under 2 weeks (usually too fast) and those that responded in over 16 weeks (usually too slow). The speed was not the differentiator; the decision-making timeline was.
Specifically, the successful responses made their key decision (respond, reframe, or pivot) within 48 hours of signal detection, then spent 6–12 weeks on execution. The unsuccessful responses either decided too fast (in under 24 hours, without full context) or delayed the decision (more than 2 weeks to decide). The execution timeline mattered less than the decision timeline.
Pattern 2 · The response moved the frame, not the feature
Successful responses rarely involved shipping a matching feature. In 28 of the 34 successful responses, the company's response was a reframe (marketing narrative, positioning shift, or campaign) rather than a feature match. In the 55 unsuccessful responses, 41 involved a feature match — and 32 of those 41 shipped the matching feature too late to matter.
The pattern: feature matches are usually losing moves in competitive responses because the competitor has the first-mover advantage and you're working off a slower cycle. Reframes are usually winning moves because they shift the competitive ground rather than fighting on the ground the competitor chose.
Pattern 3 · The response was communicated by one voice
Successful responses were communicated through one primary voice — usually the CEO or CMO, consistently across channels. Unsuccessful responses involved multiple voices (CEO, CRO, VP of Product, PMM) each communicating slightly different versions of the response. The inconsistency cost the response its market impact.
In 31 of the 34 successful responses, the external communications (blog, email, press, internal all-hands) were signed by or attributed to a single leader. In 47 of the 55 unsuccessful responses, multiple senior leaders issued separate external communications on the same response. The fragmented voice produced a fragmented market reaction.
What the archive cannot answer
Three things the archive genuinely cannot tell you, despite the volume.
The reverse-causation question. Competitors whose shifts correlate with outcomes may have been signaling a position the market was already going to move to. The archive does not distinguish "the competitor caused the shift" from "the competitor read the shift coming."
The tail of rare events. With 10,400 shifts, rare events (major acquisitions, category collapses) are still represented by 20–30 cases. Statistical inference on small samples is unreliable. The archive's findings are most reliable for high-frequency signal types.
The category-specific variations. Categories differ. The 73% careers-page prediction rate is the cross-category median; in fast-moving categories it's higher, in slow categories it's lower. Operating a monitoring program from the archive's aggregate without adjusting for your specific category will mislead you.
How to use this data
Three operational changes most monitoring programs should consider, given the findings:
Reduce feature-launch weight. Feature launches are the highest-volume, lowest-signal monitoring activity. A program that deliberately under-weights them produces a team with more time to invest in higher-value signal types.
Elevate careers-page monitoring. Most programs treat careers-page monitoring as a secondary activity. The archive data suggests it should be primary. Weekly review of tier-A competitor careers pages, with explicit pattern detection, is a high-ROI discipline.
Commit to one-voice response discipline. When responses happen, assign one communication lead. Multiple senior leaders communicating separately on the same competitive response produces the fragmentation that predicts unsuccessful outcomes in the archive.
The archive is not a prescription. It's a mirror — an opportunity to see whether your monitoring program matches the patterns that worked in 214 other companies. Most programs will find they need to reallocate, not overhaul. The reallocation is usually the difference between a monitoring program that pays for itself and one that drains PMM hours without compounding into competitive advantage.
Competitor Signals
Know what your competitors are doing before your reps find out in a deal.
Competitor Signals monitors your named competitors' public surfaces daily — pricing pages, messaging, job postings, and more — and flags the moves that actually demand a response. No noise, no Google Alerts, no manual checking.
- ✓Daily monitoring of competitor positioning moves
- ✓Filters noise from material changes
- ✓Recommended responses grounded in your own strategy
One sharp B2B marketing read, most Thursdays.
Practical frameworks, competitive teardowns, and field observations across positioning, messaging, launches, and go-to-market. Written for working CMOs and PMMs. No listicles. No vendor roundups. Unsubscribe whenever.
Keep reading
The 6 Types of Competitor Signals You Need to Track
Most monitoring dashboards track the wrong thing — they count alerts. The six signal types below are what actually moves deals, and each has a distinct cadence, owner, and response shape.
Competitor Signal Types Ranked by Threat Level
Not every competitor signal deserves the same response. Twelve signal types ranked from highest threat to background noise — with the specific response each warrants and the ones most teams over-react to.
How to Monitor 10 Competitors in 15 Minutes a Day
A weekly 15-minute review across ten competitors and three surfaces — the discipline that keeps it from becoming a two-hour time sink, plus a graduation path.