The Leadership Letter

Real correspondence from the people running real companies — and what it reveals about leadership.

Admit the Threat You Missed, Then Change Course

Strategy reviews are worthless if leaders won't say out loud what they got wrong.

My interpretation of our two-door strategy meant our primary mission was making DFP the best total revenue management solution, across direct and indirect inventory together; and AdX for non-DFPs was an edge case, not the focus. If we want YM / AdX to make sense entirely without DFP, it would mean minimizing YM features in DFP, to avoid confusion or duplicated effort.

Summarizing -- it sounds like the upshot of this strategy is to compete with the yield managers by investing in more publisher-centric features for AdX, plus an expanded services layer. The devil's advocate question I could imagine stakeholders asking is, how is this materially different from what the AdX publisher strategy was before we started talking about the YM challenge back at the strat summit?

I think getting backfill revenue through the combo of DFP and AdX was a fundamental component of our strategy (and one of the reasons why Google bought DCLK in the first place). And in general it is working — just look at the backfill growth rates. But the reality is we missed the YM threat — both on the AdX side as well as the DFP side. And this was a core mission of AdX. So I think we are altering our approach on both AdX as well as DFP and sales/services to get this right.

1. Core Message

Neal Mohan is reconciling Google's ad tech strategy after realizing the team underestimated a competitive threat from "yield managers." He restates the "two-door strategy": DFP is the main product for managing both direct and indirect ad inventory, while AdX-without-DFP is treated as an "edge case." He then admits, plainly, that "we missed the YM threat — both on the AdX side as well as the DFP side," and says the approach to AdX, DFP, and sales/services is being altered to fix it.

2. What the Executive Is Really Thinking

Mohan is doing two things at once. First, he is clarifying priorities so the team doesn't build duplicate yield management features in both DFP and AdX. Second, he is pre-empting a tough internal question — the "devil's advocate" one — about whether anything has actually changed since the strategy summit. His answer: yes, because the team now recognizes a threat it had previously missed. He also reminds the team why this matters historically — backfill revenue through DFP + AdX was "one of the reasons why Google bought DCLK in the first place." The original thesis of the DoubleClick acquisition is on the line.

3. Key Management Lessons

Name the threat you missed, out loud

What it means

Mohan writes directly: "we missed the YM threat." No spin, no softening.

Why it matters

Teams cannot re-plan around a problem leadership won't acknowledge. Naming the miss gives everyone permission to change direction without losing face.

MBA Perspective

This is basic Competitive Strategy hygiene. If a rival category (yield managers) is taking share in a job your product was supposed to own, denying it just delays the response.

Real-world application

When a competitor wins a deal you expected to win, write down what capability they had that you didn't. Circulate it. Don't bury it in a QBR slide.

Pick the primary product, make the other one the edge case

What it means

Mohan frames the choice clearly: DFP is the "primary mission"; AdX for non-DFP publishers is "an edge case, not the focus." If you change that, you have to "minimize YM features in DFP, to avoid confusion or duplicated effort."

Why it matters

When two products in the same company can solve the same problem, engineering effort splits, the sales pitch gets muddy, and customers get confused.

MBA Perspective

Platform Strategy: a platform wins by being the integration point. If DFP is the platform, AdX features should reinforce it, not compete with it.

Real-world application

If you have two SKUs or two teams chasing the same customer job, force a written decision about which is primary. The other one's roadmap should be subordinated, not parallel.

Anticipate the devil's advocate before the meeting

What it means

Mohan literally writes out the hardest question stakeholders will ask: how is the new plan "materially different" from the old one?

Why it matters

If you can't answer that question, your "new strategy" is a relabel. Writing the objection forces you to find a real answer or admit there isn't one.

Real-world application

Before presenting any pivot, write the one-line skeptical question your sharpest board member would ask. Answer it on slide two.

Tie current decisions back to the original acquisition thesis

What it means

Mohan reminds the group that DFP + AdX backfill was "one of the reasons why Google bought DCLK in the first place."

Why it matters

It anchors the team in why the strategy exists. If yield managers break that backfill loop, the original M&A rationale weakens.

MBA Perspective

Build vs Buy: the value of an acquisition is realized over years through the integrated workflow it enables. Defending that workflow against new entrants is part of the deal's ROI.

Real-world application

When reviewing a unit acquired years ago, re-read the original deal memo. Ask whether the threats the deal was meant to solve still match today's market.

Distinguish "working" from "winning"

What it means

Mohan notes backfill "is working — just look at the backfill growth rates." But growing numbers can coexist with a missed threat.

Why it matters

Growth metrics can mask competitive erosion at the edges. A product can be growing and losing strategic position at the same time.

Real-world application

Pair every growth KPI with a competitive KPI — win rate vs. a named rival, share of new logos, feature parity gap.

4. Strategic Analysis (MBA Style)

Competitive Strategy

The response to yield managers is to "compete... by investing in more publisher-centric features for AdX, plus an expanded services layer." Product depth plus services — a classic enterprise move when a point solution starts peeling off your customers' attention.

Risk Analysis

The core risk is that yield managers slot themselves between publishers and AdX, weakening the DFP + AdX backfill loop that justified the DoubleClick acquisition. A secondary risk is internal: building yield management in both DFP and AdX creates "confusion or duplicated effort."

Build vs Buy Analysis

The document only signals "build" — more features, more services. It does not mention acquiring a yield manager. Whether that was considered is not in the document.

Market Dynamics

The publisher ad tech stack in 2010 had multiple layers — ad server (DFP), exchange (AdX), and yield managers sitting on top. Mohan's framing shows Google saw yield managers as a layer that could capture publisher loyalty and decision-making, which is strategically dangerous for whoever sits below.

Long-Term Strategic Implications

If the integrated DFP + AdX approach wins, Google keeps the publisher relationship and the backfill economics. If yield managers win the publisher relationship, AdX risks becoming one demand source among many — a commoditized position.

5. Hidden Insights

  • The two-door framing is itself a tell. Calling AdX-without-DFP an "edge case" reveals that the real strategic asset is the bundled workflow, not either product alone.
  • Backfill is the moat. The phrase "backfill revenue through the combo of DFP and AdX was a fundamental component of our strategy" suggests the integration — not the exchange alone — is what Google is defending.
  • Self-criticism is targeted. Mohan says the miss was on "both... AdX as well as the DFP side" and that yield management "was a core mission of AdX." He is locating accountability inside his own org, not blaming the market.
  • The services layer matters. Mentioning "sales/services" alongside product changes hints that the fix isn't purely technical — publisher relationships and account coverage are part of the competitive response.
Court Exhibit
United States v. Google LLC (Ad Tech)
1:23-cv-00108 (VAED), Trial Ex. PTX0060 — DOJ public archive
October 19, 2010
Public domain
View the primary source →