I've seen a few threads here asking about monitoring competitors, but most focus on social listening or web scrapers. If you're using tl;dv for sales or investor calls, you're sitting on a more direct signal: spoken mentions. The platform's keyword alert feature is theoretically the right tool, but the implementation has nuances that aren't covered in their docs. A naive setup will give you a noisy, useless feed.
Here's a step-by-step based on configuring this for three separate data teams over the last quarter. The goal is actionable alerts, not just notifications.
**First, understand the data model.** A tl;dv "keyword alert" triggers on a *transcript phrase* within a *recorded meeting*. This seems simple until you realize:
* It matches substrings unless you use quotes, leading to false positives.
* It has no native concept of stemming (e.g., "CompetitorCo" vs "CompetitorCo's").
* The alert fires per meeting instance, not per unique mention.
**Step 1: Building a Clean Keyword List**
Don't just input "OurCompetitor Inc." You need a list of all verbal variations. For each competitor, create a structured list. Example for a competitor named "StratFlow":
```yaml
primary_names:
- "StratFlow"
- "Strat Flow"
- "StratFlow's"
common_mispronunciations:
- "Stratflo"
- "Strat Flo"
product_specific:
- "FlowAnalytics" # their main product
- "SF Analytics"
acronyms:
- "SF"
- "SFA"
```
You'll feed these into tl;dv as separate alert rules, grouped by competitor. Use **exact phrases in quotes** for multi-word names to avoid substring matches. For example, `"Strat Flow"` will not trigger on "We need a strategic flow."
**Step 2: Setting Up the Rules in tl;dv**
Go to Alerts -> Keyword Alerts -> Create. The critical choices:
* **Rule Scope:** Set this to "All team meetings" only if you have a centralized GCal. Otherwise, use "Specific calendars" to avoid noise from internal strategy sessions.
* **Keyword Logic:** Use "Any of these words" for your list of variations. Input them as a comma-separated list, each phrase in double quotes.
```
"StratFlow", "Strat Flow", "FlowAnalytics", "SF"
```
* **Exclude Words:** This is your most powerful filter. Add your own company name, common false triggers, and generic terms. For example: `"our flow", "strategy flow", "safe"`.
**Step 3: Routing & Triage**
The native email/Slack alert is a starting point. For a production system, use the Zapier/Make.com integration to send the alert payload (meeting title, timestamp, transcript snippet) to a data pipeline. I pipe them into a small Snowflake table with fields for:
* competitor_detected
* meeting_context
* customer_name (parsed from calendar event)
* mention_timestamp
This allows for weekly aggregation dashboards and deduplication.
**Pitfalls & Cost Notes:**
* Each keyword variation counts toward your plan's alert limit. If you have 5 competitors with 10 variations each, that's 50 alerts. On the Pro plan (50 alerts), you're already at capacity.
* The transcript snippet provided is often too short for context. You must click into the meeting and expand the transcript segment to understand the sentiment (complaint, evaluation, partnership talk).
* There is no native negative keyword support (e.g., "NOT internal"). You must rely heavily on the exclude list.
The outcome of this setup shouldn't just be a pile of notifications. It should feed a table that answers: "Which prospects are discussing which competitors, and in what frequency, over time?" That's how you move from passive listening to active competitive intelligence.
—davidr
—davidr