YouTube Analytics

YouTube Analytics is now live in LDOO. See what it unlocks

Features09 / Memory & context

LDOO is conversational analytics for agencies and marketing teams.

Answer #20 is better than answer #1.

Most analytics tools forget everything between sessions. LDOO accumulates context — baselines, conversations, goals, corrections — and actively learns from it: detecting trends, tracking recommendation outcomes, recognising seasonal patterns, correlating anomalies across clients. The AI doesn't just know the numbers. It knows what's normal, what changed, and what to do next.

What LDOO remembers

Six layers of context in every answer.

Each layer is injected into the AI prompt automatically. You don't configure any of this — it builds over time as you use LDOO.

1

90-day baselines

Statistical normals computed from the last 90 days of data for each client, source, and metric. When LDOO says "CPA is above normal," it means it — the baseline is calculated, not guessed.

2

Conversation history

Key findings from your last 30 days of conversations about each client. LDOO references prior analysis instead of starting from scratch every time.

3

Client goals & KPIs

Industry, primary KPI, targets, seasonality notes. When LDOO says "ROAS is within your target," it's checking against goals you set — not making a generic judgment.

4

Report memories

After every report is generated, LDOO distills the narrative and recommendations into topic-tagged memories. Future answers reference these — efficiency, channel focus, known issues.

5

Active alerts

Unread anomaly alerts are injected into the prompt so LDOO can proactively surface them. "CPA spiked on Tuesday" appears in the answer before you ask about it.

6

Learned corrections

Thumbs-down an answer or correct something in conversation, and LDOO records it as a durable fact. "Attribution is 7-day click" or "Brand NZ was paused" — the same mistake never repeats.

How baselines work

Computed, not configured.

After every data sync, LDOO recomputes 90-day statistical normals for each client, source, and metric. This is what powers the difference between “CPA is $65” and “CPA is $65 — above your normal $40-$55 range and the highest in three months.”

Per-client, per-source
Baselines are scoped to each client and each data source. What's normal for one client isn't normal for another — LDOO knows the difference.
Auto-recomputed
Every time new data syncs, baselines update. No manual configuration. The system stays current without you touching it.
Deviation detection
When a metric moves more than 20% from baseline in 7 days (or 15% in 30 days), LDOO flags it as unusual — and can surface it as a briefing card on app open.
Grounded recommendations
"Increase budget by 15%" carries more weight when LDOO can show that the current CPA is 30% below the 90-day average. Baselines make recommendations specific.
What LDOO learns

Context is the input. Intelligence is the output.

Beyond remembering what you tell it, LDOO actively learns from data patterns and outcomes. These signals build automatically with use — no configuration needed.

Multi-month trends
When a metric moves in the same direction for 3+ consecutive months, LDOO detects it and weaves it into answers and reports. "This is the 4th consecutive month of organic growth."
Seasonal patterns
After 12 months of data, LDOO identifies calendar-driven shifts — Q4 CPA spikes, summer traffic dips — and warns you 30 days before the season arrives.
Recommendation outcomes
Every recommendation is tracked against the metric it targeted. Did pausing that campaign actually reduce CPA? LDOO checks and tells you — then uses the outcome to improve future advice.
Cross-client correlation
When the same anomaly hits 3+ clients on the same platform within 24 hours, LDOO tells you once: "likely platform-wide." Redundant per-client alerts are suppressed automatically.
The difference

Without context vs with context.

This is the same question asked of the same data. The difference is whether the AI has accumulated context about the client.

Without context

“CPA this week is $65. This is higher than last week's $44. Consider reviewing your campaigns and adjusting bids to reduce costs.”

With LDOO context

“CPA is $65 — above your normal $40-$55 range and the highest in three months. The spike is concentrated in the Retargeting campaign where frequency hit 6.2. Last time this happened (Feb 12), refreshing creative brought CPA back to $42 within a week. Your target is $50, so this needs action.”

Key details

Always on. Zero configuration.

Baselines
Auto-computed

Statistical normals recomputed after every sync. 90-day window. Per client, per source, per metric.

History
30 days

Key findings from the last 30 days of conversations are injected into every answer. Prior analysis compounds.

Trends
3+ months

Multi-month directional trends detected automatically. Woven into answers and reports as momentum signals.

Seasonal
12+ months

Calendar patterns detected after a year of data. Briefing cards warn you 30 days before a seasonal shift.

Recommendations
Tracked

Every recommendation is checked against outcomes at metric-appropriate intervals. Results feed back into future answers.

Visible to you
Badge on answer

"Remembering N prior conversations" badge on every answer card. You always know what context the AI is using.

FAQ

Client memory questions

What LDOO remembers, how baselines work, and why answers get sharper the longer you use it.

Six layers: 90-day statistical baselines, conversation history from the last 30 days, client goals and KPI targets, report memories distilled from prior reports, active unread alerts, and learned corrections from your feedback.

The AI that knows
your business.

Every question, correction, and report makes the next answer better. Context accumulates. Intelligence compounds. By month 3, the difference is measurable.