Verified against Perplexity Pro · 2026-07-02
Map a trend across time with a dated citation for every data point, not a snapshot
A trend-analysis prompt that requires a chronological, dated citation table and an explicit split between a real structural shift and a one-off headline dressed up as a trend.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Analyze this as a trend over time, not a single current snapshot. Every data point needs its own date and citation — a trend claim with no dates attached is not a trend claim. TREND TO ANALYZE Adoption of usage-based pricing among B2B SaaS companies TIME WINDOW 2021 to present MARKET / GEOGRAPHY North American mid-market SaaS METRIC TO TRACK % of new SaaS pricing pages listing usage-based tiers as the default option OUTPUT FORMAT 1. A chronological table: Date | Data point or event | Source. Sorted oldest to newest. 2. A short narrative distinguishing a real structural shift — sustained direction backed by multiple dated points — from noise: a single headline, a one-off event, or a seasonal blip that looks like a trend but isn't. 3. Flag explicitly if the most recent one or two data points might be too new to confirm the trend is continuing, rather than extrapolating confidently off a single recent point. 4. What would have to happen for this trend to reverse, based on what the sources say is actually driving it.
Customize the highlighted detailsoptional — the prompt above already works
Why this works
Asked for a trend without structural constraints, a model tends to answer with the current state framed as if it were the whole trend, because the most recent and most abundant sources are about now, not about the shape of change over time. Forcing every row of the output to carry its own date converts an impressionistic 'this is growing' into a chronological table a reader can actually eyeball for direction, gaps, and inflection points. The explicit instruction to separate structural shift from noise directly counters a specific and common error — treating a single recent headline as proof of a trend, when trend claims require multiple dated points moving the same direction, not one data point plus a narrative. Flagging the newest points as potentially unconfirmed also guards against the model's tendency to extrapolate a clean line through the most recent, least-verified data, which is exactly where a trend narrative is most likely to be wrong.
Verified against
Perplexity Pro Sonar Pro · 2026-07-02
Changelog
- 2026-07-02 — Initial publish, verified against Perplexity Pro Sonar Pro search.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
