Writing

Essays and short responses on technology, investing, operations, and media.

Archive

Essays and responses since November 2025.

August 2026

  1. Ask for the lost deals before the reference list

    The first diligence list worth asking for contains the deals the company lost, plus the customers that renewed flat.

  2. Underwrite the pro rata like a first check

    A fully available pro rata can hide financing risk, so write the follow-on memo at the new price and set reserve policy outside a live round.

July 2026

  1. Channel partnerships are decided in the comp plan

    A reseller agreement sets margin and territory. Quota credit and billable hours decide whether the partner's sellers ever carry the product.

  2. Write the earnout on what the acquired team still controls

    Post-close pricing, staffing, and cost allocation can move an earnout. Tie the metric to work the acquired team still runs, then write the buyer's constraints into the contract.

  3. A free pilot is research the seller pays for

    Free enterprise pilots without a named deployment budget are seller-funded research. Keep them out of qualified pipeline until a recurring cost center has an owner.

  4. AI can make the backlog worse

    Cheap output moves the bottleneck to review, decisions, and integration. Hours saved mean little if the organization cannot absorb the work.

  5. Generating the sports clip will be the cheap part

    The defensible layer in sports video is the rights and delivery system around the clip, not the act of finding the moment.

June 2026

  1. Sixty hours of agent work a day is a management failure

    Parallel agents can create more work than one person can review. The bottleneck moves from generation to acceptance.

May 2026

  1. Staying private can hide a weak operating cadence

    Public readiness forces mature companies to explain their economics, close their books, and accept price discovery.

  2. Enterprise AI should attack waiting, not headcount

    The strongest enterprise use case is compressing handoffs, approvals, and queues instead of converting recovered minutes into a layoff model.

March 2026

  1. Operator-investors should operate less

    Operating experience helps investors when it improves a founder's judgment. It becomes a liability when advice turns into shadow management.

February 2026

  1. Sacramento does not need another accelerator

    The region's next entrepreneurship dollar should recruit repeat operators and buy customer access instead of adding another pitch cohort.

January 2026

  1. Every integration plan needs a no-integration list

    Acquirers should name the systems and practices they will leave alone until they understand what creates the target's value.

December 2025

  1. Your first U.S. hire should not be a salesperson

    Complex companies entering America need a customer operator before a quota carrier because delivery creates the repeatable sales motion.

  2. Open standards will erase most agent moats

    As agent connectivity standardizes, connector counts lose value. The durable work moves to permissions, exceptions, and recovery.

November 2025

  1. Private equity wants your 401(k)

    Broader access to private markets may solve fund managers' distribution problem before it improves retirement portfolios.

  2. Management latency breaks growing companies

    Growing companies slow down when decision rights lag headcount. Tracking the delay exposes where authority stopped scaling.