The Yield Doctor with James Deaker

What happens to ad tech business models when AI agents, not humans, are running the queries? Shay Brog, CEO of Burt Intelligence, joins me to talk about building the operational intelligence layer for ad-supported media, where automation is paying off for publishers right now, why yield decisions still need an accountable human, and whether SSPs should move to a SaaS model.

This conversation is part of my series on the companies driving AI change for publishers and ad tech.

Chapters

0:00 Introduction

0:35 Shay's path: Jerusalem, Sydney, New York

1:25 The origin of Burt

2:39 Why Shay joined: publisher-facing, not smoke and mirrors

5:04 What Burt does: the operational intelligence layer

6:54 SaaS or managed service?

9:39 The semantic layer: 90% standard, 10% custom

12:46 Investing in AI and MCP

15:37 Human in the loop and accountability for yield decisions

20:30 Where automation pays off: reconciliation, client reporting, forecasting

25:05 Grow revenue or cut costs?

27:52 AI costs, token pricing, and should SSPs be SaaS?

32:13 Contrarian view: optimism

35:00 Where to find Shay and Burt

Shay Brog on LinkedIn: [add URL]

Burt Intelligence: https://www.burtintelligence.com/

I'm James Deaker, The Yield Doctor. Through Korukea Media, I help publishers and ad tech companies with pricing strategy and yield management. https://korukeamedia.com/

Subscribe to The Yield Doctor Report: https://korukeamedia.com/newsletter/

How do you get one ad ops person doing the work of 15 to 20? Frans Vermeulen, President of Swivel, explains how AI-driven automation is changing yield management, programmatic optimization and ad operations for large publishers and CTV broadcasters, without ripping out the ad stack.

Frans has spent 25 years in digital advertising at DoubleClick, FreeWheel, Comcast, Tru Optik and TransUnion. In this conversation we cover:

- Why Swivel is an orchestration layer on top of existing ad servers, not another ad server

- How natural language creates the rules, while deterministic scripts make the changes

- Why automating the 900 small optimizations a human team never gets to adds up to real incremental revenue

- How to prove revenue uplift with holdback and control groups

- The "Rosetta Stone" problem of fragmented ad platforms, APIs and data latency

- Why the UI isn't dead, and agent-to-agent trading won't take over the world

- Where the CTV marketplace is heading as supply grows and prices come under pressure

Learn more about Swivel: https://www.swivel.ai

Chapters:

0:00 Introduction: AI in the ad tech landscape

1:51 Swivel's origin story and the SpringServe team

6:21 Why ad ops problems never seem to get solved

7:58 What Swivel does: scaling ad operations

12:32 Where Swivel sits in the ad stack

16:17 Where Swivel innovates: AI plus deterministic rules

20:43 Replacing human actions vs. doing what humans can't

22:33 Proving revenue uplift with holdback groups

25:05 Fragmented systems, data latency and the "Rosetta Stone"

27:34 Giving power back to the sell side

29:27 Contrarian views: the UI isn't dead

33:22 The state of the CTV marketplace

36:28 Where to find Frans and Swivel

I'm James Deaker, The Yield Doctor. Through Korukea Media, I help publishers and ad tech companies with pricing strategy and yield management. https://korukeamedia.com/

Subscribe to The Yield Doctor Report: https://korukeamedia.com/newsletter/

What happens when an analytics company spends nine months teaching an LLM to understand ad revenue data? I talk with Nico Schueller, CEO and co-founder of Adomik, about how the company went from floor optimization on AppNexus to JP, its AI infrastructure for publishers.

We cover:

What’s inside Adomik’s market data set, and why accuracy matters more than completeness

Why JP is infrastructure, not an agent or a chatbot

How a semantic model makes LLM analysis of transactional data reliable

Onboarding speed, and JP’s December release with versions for AdOps, sales and yield

Whether publishers can trust their data again

When to keep a human in the loop

Nico’s contrarian view that enterprise AI belongs in enterprise software companies

This conversation is part of my series on the companies driving AI change in digital media and ad tech.

Chapters

0:00 Why I’m talking to AI-driven ad tech companies

1:29 Adomik’s origin story: AppNexus and floor optimization

3:52 Header bidding and the shift to revenue management

4:45 Building the market data set

5:35 AI and the launch of JP

6:37 What’s inside Adomik’s market data set

13:34 Directional or reliable? Completeness vs. accuracy

16:01 JP is infrastructure, not an agent

17:41 Why the semantic model matters

21:25 Revenue Analyst, Market Analyst and JP as an MCP server

22:48 How long does onboarding a new client take?

26:17 JP’s December release: JP for AdOps, Sales and Yield

28:05 Can publishers trust their data again?

31:47 Guardrails and the human in the loop

35:29 Nico’s contrarian view: enterprise AI is enterprise software

39:43 Where to learn more about Adomik

Learn more about Adomik: https://www.adomik.com/

Connect with Nico on LinkedIn: https://www.linkedin.com/in/nschueller

Article on this conversation:

https://korukeamedia.com/enterprise-ai-is-enterprise-software-what-adomik-learned-building-jp/

I’m James Deaker, The Yield Doctor. Through Korukea Media, I help publishers and ad tech companies with pricing strategy and yield management. https://korukeamedia.com/

Subscribe to The Yield Doctor Report: https://korukeamedia.com/newsletter/

Title: Bonus Impressions vs. Price Discounts: Which Protects Publisher Yield?

Description:

When an advertiser asks for a concession on a $10 CPM, should you drop the rate to $8 or give them 25% more impressions? Mathematically, a $10,000 budget gets the same result. But from a yield management perspective, they behave very differently.

In this video, I break down the mechanics of price discounting versus added value (bonus inventory). We look at why dropping your floor prices creates dangerous anchoring effects, how bonus inventory can protect your perceived value (while introducing its own operational headaches), and how programmatic demand completely changes the equation.

⏱️ Chapters:

0:00 - The $10,000 Dilemma: Discounts vs. Added Value

1:27 - What Buyers and Agencies Actually Prefer (And Why)

3:03 - The Danger of Price Discounts & Anchoring

4:14 - Why Bonus Inventory Protects Your Floor Price

6:14 - The Operational & Opportunity Costs of Added Value

7:44 - How Programmatic Demand Changes the Equation

9:11 - Beware of Double Dipping & Internal Economics

10:17 - Final Verdict: Which Concession Should You Choose?

🔗 Resources & Links:

Read the full deep-dive article on my blog:

https://korukeamedia.com/bonus-impressions-vs-price-discounts-which-protects-publisher-yield/

What is Kana, and how is AI reshaping enterprise marketing? In this episode, Vivek Vaidya (Co-Founder & CTO of Kana; former CTO of Salesforce Marketing Cloud; co-founder of Krux and Habu) breaks down Kana's vision, tech stack, and how they are automating marketing admin to accelerate growth.

Watch to learn how Kana intelligently structures enterprise context, where AI agents actually deliver ROI, and what the future of data-driven customer engagement looks like from one of enterprise tech's most experienced architects.

🔔 Subscribe for in-depth breakdowns and interviews with leading digital media and tech founders: https://www.youtube.com/@jamesdeaker

━━━━━━━━━━━━━━━━━━━━━━━━━━

TIMESTAMPS / CHAPTERS

0:00 - Introduction & Welcome

01:30 - Who is Vivek Vaidya? (Rapt, Krux, Salesforce, Habu)

04:15 - What is Kana? Solving the Marketing Admin Problem

08:40 - How Kana's Architecture Integrates with Enterprise Systems

13:10 - The Role of AI Agents & Context Structuring

18:25 - Overcoming Legacy Enterprise Data Bottlenecks

23:50 - The Future of AI in Marketing and Customer Engagement

28:10 - Key Takeaways & Where to Learn More About Kana

━━━━━━━━━━━━━━━━━━━━━━━━━━

ABOUT VIVEK VAIDYA

Vivek Vaidya is a serial entrepreneur, technologist, and Co-Founder/CTO of Kana. Over a 25+ year career, he has built and scaled industry-defining enterprise platforms, including Rapt (acquired by Microsoft), Krux (acquired by Salesforce), and Habu (acquired by LiveRamp). He previously served as CTO of Salesforce Marketing Cloud and is a founding General Partner at super{set}.

RESOURCES & LINKS

• Learn more about Kana: https://www.kana.com [or your preferred link]

• Connect with Vivek Vaidya: https://www.linkedin.com/in/vivekvaidya/

• Connect with [Host Name / Company]: [YOUR_LINKEDIN_OR_WEBSITE]

• Sign up for our newsletter: [YOUR_NEWSLETTER_LINK]

#MarTech #AdTech #EnterpriseAI #Kana #MarketingAutomation #TechInterview

Struggling with publisher direct sell-through rate (STR) targets? Whether you're managing direct sales inventory or retail/commerce media networks, treating STR like an airline load-factor quota destroys revenue.

In this video, James Deaker (The Yield Doctor) breaks down why 100% sell-through creates operational stress, why anchoring to 75/85/90 targets fails, and how to use the 4-part diagnostic framework instead.

Chapters:

0:00 - Introduction & The Sell-Through Trap

1:16 - Why STR Comes Up (And Why It Doesn't Work Like Airlines)

3:03 - The Goldilocks Metric & Examining the Extremes (0% vs. 100%)

5:32 - What Determines a Healthy STR Range?

7:13 - The Danger of Anchoring (Especially in Commerce Media)

8:54 - Changes Over Time: Why Trends Matter More Than Absolute Numbers

10:20 - The 4 Questions You Should Ask Instead of Targeting STR

Key Concepts Covered: Inventory forecasting, direct vs. programmatic channel conflict, rate card protection, and publisher yield optimization.

The FTC alleges that Amazon has been representing that it's ad marketplace runs a second priced auction, when in practice it is essentially running first priced.

There are a number of aspects of the story that have been overlooked, including the fact that Amazon could have moved to a First Priced Auction like the rest of the Display advertising marketplace.

This week it was my pleasure to join Geoff Wolinetz and Greg MacDonald on the @LeadershipIn podcast where we had a three-way roundtable about our experiences as independent consultants.

We discuss:

* The pros and cons of being solo vs. growing a team

* How to think about the models of advising vs. fractional vs. workforce augmentation

* The headwinds and tail winds for an independent consultant from the rise of AI solutions

During the conversation we independently called out the usefulness of the resources provided by David A. Fields

This episode is for anyone thinking about making the leap into independent consulting.

What do you do when leadership says, “We need next quarter’s rate card pulled together fast”?

The natural instinct is often to react to the last lost deal or competitor price drop—but updating a rate card under pressure requires focusing on the signals that actually move the needle.

In this video, we break down a structured, step-by-step approach to running a regular rate card update without starting from scratch or reinventing the wheel.

In this video, you’ll learn:

The 3 Core Questions: How to balance backward-looking performance, external market trends, and internal product changes.

Key Data Inputs: What metrics matter most—from price realization and discounting to deal review exceptions and projected sell-through.

Aligning with Finance & Sales: How to turn your data into a clear assumptions list before touching a single price.

The 3 Outcomes: How to determine whether to leave your rate card as-is, make targeted pricing tweaks, or plan a full structural overhaul.

There are areas where you need to careful of using AI to help in Pricing and Yield in Digital Advertising. It will give you an answer, but without the right oversight that answer may hurt your business.

⏱️ Timestamps

[00:00] - Appropriate Use Cases for AI in Pricing

[01:55] - Trap #1: Algorithmic Competitor Price Setting & Collusion Risk

[03:35] - Trap #2: Relying on AI for Real Pricing & Product Innovation

[06:18] - Trap #3: Designing Pricing Before Defining Your Objective Function

[08:41] - Key Takeaway: Executing Strategy vs. Defining Strategy

💡 Key Takeaways

Where AI Belongs: Executing clear internal policies, enforcing deal approval thresholds, and accelerating analytics.

Algorithmic Collusion: Why allowing AI agents to continuously monitor and react to competitor pricing creates unintended legal and operational gray areas.

Innovation Bottlenecks: Why AI adapts existing models (CPMs, sponsorships) well, but human insight is needed for breakthrough models like Uber's "Cost Per Ride" (CPR).

The Objective Function: Why asking AI to optimize pricing before management explicitly defines what success looks like yields great answers to the wrong questions.

📌 About The Yield Doctor

Hosted by James Deaker (The Yield Doctor), this channel delivers practical yield frameworks, ad tech strategy, and AI governance insights for media and digital advertising executives.

🔔 Subscribe for practical yield strategy & AI risk controls

🌐 Connect & Discussion

Are you seeing companies hand over pricing decisions to AI that they shouldn't? Let us know in the comments!