Actionable insights — Guest Thesis: Zeta Global

2:23 1. Screen for price drops that the fundamentals don't explain

1. Run a screener that flags large share-price declines while revenue, margins and guidance are unchanged.
2. Park the name on a watchlist; don't buy the headline (often a short report).
3. Wait for the next earnings call / independent evidence (an auditor's review) that addresses the specific allegation, then start the real research.

• Whether the residual allegation is fundamental (the product doesn't work) or definitional/legal (what counts as "opt-in").

1:07:03 2. Model the growth you get from customers you already have

1. Find the share of each client's total category budget the company captures today (wallet share).
2. Take management's guided annual increase; adjust only if there is a documented sandbagging record.
3. Project revenue with zero new customers; compare that path with the analyst consensus.
4. Cross-check against net revenue retention (above 100% = existing customers spending more).

• Wallet share flat for a year — the guest's own red flag.
• NRR trending back toward 110.

25:15 3. Ask who owns the data pipe, not just the model

1. List where the company's signals come from: owned properties, licensed third-party data, or partners.
2. Owned, consented collection points are a moat; leased data is available to every competitor.
3. Check how the client's own data is treated — isolation from other clients is what makes enterprises upload it.

• Privacy rules that reclassify how consent is obtained (cookie/login/newsletter opt-ins).

1:02:11 4. Use a proven analog as a story, not a price target

1. Name the company whose playbook the candidate resembles and the specific overlap (value proposition, customer, model).
2. List the differences explicitly.
3. Ask what return you need, not whether it repeats the analog — the catch-up name doesn't need to match the leader's gain.

• The analog doing the valuation work: if the pitch only holds at a Palantir multiple, it doesn't hold.

1:40:59 5. Write down the kill criteria before the position works

1. Identify the two or three engines the thesis actually depends on.
2. For each, set an observable metric and a failure level.
3. Revisit each quarter; sell on the metric, not the price.

• Organic growth ex-acquisitions; Palantir-sourced deal count; super-scaled customer ARPU.

1:33:50 6. Set both a buy zone and a trim zone for a concentrated holding

1. Record where you accumulated and why that price offered a margin of safety.
2. Name a price at which you'd reluctantly trim even while still bullish.
3. When a great company runs past fair value, trim a slice to fund better valuations rather than exit.

• Position weight drifting far above plan (42% here) as the stock re-rates.