Refresh cadence and pipeline schedule

Documentation Index

Fetch the complete documentation index at: https://docs.getcited.in/llms.txt

Use this file to discover all available pages before exploring further.

Cited’s monitoring pipeline runs on a fixed daily schedule. Understanding when data is collected, processed, and available in dashboards helps brands interpret metric movements correctly and avoid over-reacting to single-day fluctuations.

​

Daily pipeline schedule

The pipeline runs in four stages each day, timed sequentially so each stage completes before the next begins. All times are UTC. The platform queries stage is the most resource-intensive — it makes hundreds of API calls across 5 platforms for all tracked brands. The aggregation stage runs the Haiku parsing pipeline on each collected response.

​

What “daily” means in practice

Each day’s pipeline run produces one data point per brand per query per platform. Dashboard metrics — mention rate, share of voice, average position, sentiment — are computed from the aggregated results of that day’s run. Week-over-week and month-over-month trends are computed from the series of daily aggregates. For a Pro-tier brand tracking 75 queries across 5 platforms, each daily run produces 375 data points. This volume is sufficient for stable per-platform metrics but may show variance on per-query breakdowns. Two-week moving averages smooth out most non-deterministic noise.

​

Narrative and recommendation refresh cadence

Your AI Narrative profile, perception gap analysis, and prioritized recommendations regenerate on a plan-tier-aware schedule: Why twice-weekly for Pro? D2C and B2B SaaS visibility shifts can compound within days, especially during product launches or competitor activity. Twice-weekly cadence captures these shifts within 3-4 days while keeping API costs sustainable. Recommendation expiry: Recommendations expire and are replaced with fresh analysis on each regeneration cycle. Pro plan recommendations carry a 4-day TTL, Starter plan recommendations carry an 8-day TTL. Each recommendation’s Impact Score is recomputed on every refresh cycle based on the latest pipeline data. On-demand refresh: Pro and above accounts can request immediate narrative regeneration (for example, before a sales presentation or after major content launches). Contact [email protected] for now — self-serve UI coming soon.

​

Cited Index refresh cadence

The Cited Index is refreshed monthly, not daily. Each monthly edition re-runs all 185 queries across 253 brands on 4 platforms — 740 LLM calls plus 740 Haiku parsing calls, totaling 1,480 API calls per refresh. Cost per refresh: approximately $10. Wall-clock time: approximately 4 hours running serially. The March 2026 edition is the current published benchmark. The pipeline is triggered manually via admin API rather than on an automated schedule.

​

Why not real-time

Three practical reasons prevent real-time monitoring: Cost. Real-time monitoring at the query-per-minute level would be prohibitively expensive across 5 platforms. Even daily monitoring at Pro tier generates ~$2-3 per brand per day in API costs. Non-determinism. Single-run measurements are noisy. Daily aggregation across multiple runs produces more stable metrics than a continuous stream of single-point readings. AI platforms update on multi-week cycles. The underlying training data and retrieval behavior of AI platforms changes slowly — real-time monitoring would not capture faster signal than daily monitoring does. The bottleneck for visibility change is the AI platforms’ own refresh cycles, not Cited’s measurement frequency.

​

Data availability timing

After the nightly pipeline completes (typically by 23:30-00:00 UTC / 5:00-5:30 AM IST), updated dashboard data is available within minutes. Brands checking their dashboard in the morning see the previous day’s data. There is no multi-day lag between data collection and dashboard availability. Weekly reports, generated every Monday at 00:30 UTC, summarize the past 7 days of daily data. These reports are available in the dashboard by Monday morning IST.

​

Related concepts

• Non-determinism — why daily aggregation matters
• Data freshness — how to interpret confidence in the data
• How we generate queries — what queries the pipeline runs
• Cited Index benchmarks — the monthly benchmark dataset

​

Frequently asked questions