Google Search Console

Search Console is now live in LDOO. See what it unlocks

Features09 / Memory & context

LDOO is conversational analytics for agencies and marketing teams.

Answer #50 is better than answer #1.

Most analytics tools forget everything between sessions. LDOO accumulates context — 90-day baselines, conversation history, client goals, report insights, feedback corrections. Every answer is informed by everything that came before it. The AI doesn't just know the numbers. It knows what's normal, what changed, and what you care about.

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

Feedback patterns

When you give thumbs-down on an answer, the pattern is recorded. Future answers for that client avoid the same mistake. Quality improves with use.

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.
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.

Report memory
Topic-tagged

Narratives and recommendations from generated reports are distilled into reusable memories by topic.

Visible to you
Badge on answer

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

The AI that knows
your clients.

Every question you ask makes the next answer better. Baselines, goals, history, feedback — context accumulates automatically.