Doubleverify Holdings, Inc. · NYSE:DV
Doubleverify — Investor Day (September 2023)
87 slides from Doubleverify's investor day presentation, Sep 14, 2023
Slide 1: DV Innovation Day
[D\/ INNOVATION DAY
Slide 2: Disclaimer
Disclaimer This presentation contains “forward-looking statements” that are based on management’s beliefs and assumptions and on information currently available to management. These forward- looking statements include, but are not limited to, statements about our plans, objectives, expectations and
Slide 3: Perils and Profit in the Age of AI
Perils and Profit in the Age of AI Mark Zagorski, Chief Executive Officer 3
Slide 4: We Keep Growing
We Keep Growing MFA Solution Launch New Logos Tik Tok Scibids Language Closed! Expansion 4
Slide 5: Today's Agenda
Today’s Agenda Session 1 Session 5 Session 8 Perils and Profit in Fireside Chat Panel with Dentsu the Age of AI with Diageo and The ANA Session 2 Session 9 Introducing Scibids BREAK Financial Update Session 3 Session 6 Session 10 Fireside Chat with AI @ DV: R&D Scale, Q&A (20) The Trade Desk Impact
Slide 6: AI: Perils and Profit
AI: Perils and Profit The Market Our Approach The Opportunity 6
Slide 7: New Market Challenges
New Market Challenges Content Quality Fraud Scale & Complexity 7
Slide 8: Content Quality: MFA
Content Quality: MFA By the Numbers From dozens of articles a day to over 1,000 (per site) 99% of all content could be AI generated by 2030 19% of the programmatic bidstream 8
Slide 9: Content Quality: Inflammatory News & Disinformation
Billion B Content Quality: Inflammatory News & Disinformation A Bigger Megaphone Election season risks: 25-30% increase in incendiary content Mis/dis-information generated by A.I. is more convincing Deepfakes and synthetic news 9
Slide 10: AI Driven Fraud & IVT
AI Driven Fraud & IVT By the Numbers 2X as many individual mobile app schemes YTD 137% more audio fraud schemes YTD 92%+ number of CTV fraud schemes YTD Cyclone-bot! 10
Slide 11: Scale & Complexity: Short Form UGC
Scale & Complexity: Short Form UGC New Challenges Reels, Shorts and TikTok eating the ad world 167M hours of video streamed per minute on TikTok alone 34M videos a day uploaded on TikTok Video, images, text, audio 11
Slide 12: It's Scary.
It’s Scary. 12
Slide 13: But There is Hope.
But There is Hope. 13
Slide 14: Fire, meet fire.
Fire, meet fire. 14
Slide 15: Faster, Smarter, More Efficient
Faster Smarter More Efficient 15
Slide 16: Faster: Content Classification
Faster Content Classification ~3X Faster 20%- 40% 20x Model Increased accuracy of More Cost Effective Development video classification Video Analysis models 16
Slide 17: DV Activation Evolution & Expansion
Smarter DV Activation Evolution & Expansion DV DV DV Standard Activation Dynamic Activation AI Activation Data Source Static Variable Custom Multivariate Decisioning Binary Binary Spectrum Brand : Enterprise Client Type Brand Brand : Enterprise DR : Ent + Mid Market Standard Viewabilty, ABS DV Produ
Slide 18: DV Activation Evolution
Smarter DV Activation Evolution 1 Differentiate our offering VALUE 2 Create a more sticky value proposition 3 Provide broader opportunities to leverage DV data VOLUME 4 Open an entirely new customer set 18
Slide 19: Client & Operating Efficiencies
Efficient Client & Operating Efficiencies Conversational AI tools Co-pilot Solutions for for Client Support Engineers 19
Slide 20: Where Do We Land?
Where Do We Land? @ Better Outcomes for All More Transparent, Outcomes complex, independent rule the day hazardous and AI driven solutions challenging can solve those challenges 20
Slide 21: Opportunities for Our Clients
Opportunities for Our Clients Stronger Outcomes, Safer Results 63% 48% 79% Greater attention Average decrease Average reduction using algorithmic in brand safety incidents in fraud and IVT optimization through for advertisers using incidents for advertisers Scibids using DV prebid DV prebid STRONGER
Slide 22: Opportunities for DV
Opportunities for DV Stronger & Safer Outcomes = An Essential Partner Our investments in AI driven innovation has uniquely positioned DV to: Provide more value for Solidify the essential nature customers & Move faster and more of our business to not only expand our opportunities efficiently protect
Slide 23: Protect and Perform
PROTECT IDV AND PERFORM
Slide 24: Introducing Scibids
Introducing Scibids Doug Campbell, Chief Strategy Officer 24
Slide 25: DV AI
82d.[2d.S2d.282.S81 A > (D\V | INNOVATION DAY
Slide 26: New Activation Path
New Activation Path DV Value Proposition Static Segments Algorithmic Data inside a DSP (Ex. inside Scibids Fraud, Brand Safety) (Ex. Attention) Dynamic Segments inside a DSP (Ex. ABS) 26
Slide 27: Meet Scibids
Meet Scibids Remi Lemonnier and Julien Hirth, Co-Founders 27
Slide 28: Co-Founders of Scibids
Co-Founders of Julien Hirth Remi Lemonnier General Manager President 28
Slide 29: Scibids AI Solves a Key Problem For Advertisers
Scibids AI Solves a Key Problem For Advertisers 29
Slide 30: Solving for Unique KPIs Using Unique Business Intelligence
Solving for Unique KPIs Using Unique Business Intelligence Ad Spending Business Outcomes Unique KPIs Unique Data 30
Slide 31: Solving for Unique KPIs (cont'd)
Solving for Unique KPIs Using Unique Business Intelligence (cont’d) Examples E-Commerce CPG Insurance Specific SKU Offline Customer Unique KPIs sales Sales Lifetime Value Product Share of shelf, Quality Unique Data margins Zip code, POS Scores 31
Slide 32: Diversity of Objectives, Approaches and Data
Diversity of Objectives, Approaches and Data Competitive Advantage 32
Slide 33: Address Advertisers' Needs and Complement DSPs' Offerings
ENABLED In Order to Address Advertisers’ Needs and Complement DSPs’ Offerings 33
Slide 34: Largest and Most Advanced DSPs Offer Open APIs
Largest and Most Advanced DSPs Offer Open APIs OPEN APIs BIDDING STRATEGY customizable AI informs the DSPs bidding strategy Bring Your Own Algorithm (BYOA) Bid request 1 MEDIA BUYING 2 DSP Bidder Bid response Ad Exchanger 34
Slide 35: Your DSP does what you tell it to do.
Your DSP does what you tell it to do. So tell it what you want it to do. “How to Train Your DSP”, Jounce Media 35
Slide 36: Advertisers can Improve Performance By
Advertisers can Improve Performance By 1 2 Customizing KPIs Customizing data inputs to generate real business which inform impression outcomes valuation Seamlessly using Scibids’ customizable AI 36
Slide 37: Scibids' Customizable AI
Scibids’ Customizable AI Trained to understand advertisers’ real-world business objectives Leverages intelligence across the advertisers’ business Generates bidding algorithms executed in a DSP 37
Slide 38: Scibids is independent and privacy-forward
Scibids is independent and privacy-forward 38
Slide 39: Scibids AI Unlocks Value for Advertisers and DSPs
Scibids AI Unlocks Value for Advertisers and DSPs Improvement in Performance Increase in Ad Spend +70 % % +10 % 40 +30% 39
Slide 40: Examples of Standard KPIs
Examples of Standard KPIs Performance KPIs Brand KPIs Maximize conversions Maximize viewable impressions Maximize clicks Maximize complete in-view Cost per Acquisition (CPA) and audible Click-Through-Rate (CTR) Maximize viewable for at least 10 seconds Return-on-Ad-Spend (ROAS) 40
Slide 41: Example of a Custom KPIs
Example of a Custom KPIs eCPM frequency QCPM = EFFECTIVE CPM QUALITY CPM Viewability TCCC% / Brand Safety% / (100% - IVT%) Viewable Impressions [XX% in view, X sec display, X sec video] Viewability TCCC % = /Measured Ads [XX% in view, X sec display, X sec video] Brand Safety % = Passed Ads/Total Eli
Slide 42: AI In Action for the Biggest Brands Globally
AI In Action for the Biggest Brands Globally Objective Results App registrations -60% +22% (measured by Adjust) CPC Incremental Registrations Leading CPG -52% +28% qCPM / Attention Company qCPM Authentic Attention -66% $2.5M YouTube efficiencies CPMs Media Efficiencies 42
Slide 43: Optimize Activate Automate
Optimize Activate Automate 43
Slide 44: Brand DSP
ID\V | INNOVATION DAY
Slide 45: Driving Differentiated Value on The Trade Desk
Driving Differentiated Value on The Trade Desk Cory Grever, Senior Director, Global Trading Strategy, The Trade Desk, Remi + Julien 45
Slide 46: DV-LUTION (cont'd)
DV-LUTION (cont’d) Doug Campbell, Chief Strategy Officer 46
Slide 47: Recap
ID\V | INNOVATION DAY
Slide 48: Scibids Strategic Rationale
Volume - Enables direct action 1 on DV data Value - Increases value of 2 DV performance data Scibids Strategic Rationale Velocity - Increases 3 workflow automation Vision - Acquisition of 4 a brilliant team 48
Slide 49: Data is Directly Actionable and Tailored to Our Client's Needs
1 Data is Directly Actionable and Tailored to Our Client’s Needs Segment Type Segment Type Advantages Static Segments ● Best for: Fraud, Brand Safety Data ● Binary Decisions ● Best for: ABS Data Dynamic Segments ● Custom Settings / Binary Decisions Algorithmic Activation ● Best for: Performance Data
Slide 50: DV's Granular Data Becomes More Valuable - Attention
2 DV’s Granular Data Becomes More Valuable - Attention Exposure Index Engagement Index Measurement of Ad Presentation Measurement of User Interaction • Viewable time • User presence • Video presentation • Ad interaction • Ad share of screen Based on 50+ data points Aggregate this data to create a se
Slide 51: DV's Granular Data Becomes More Valuable - Attention
2 DV’s Granular Data Becomes More Valuable - Attention Log Level Data 50 data points fed into the algorithm Conversion Data, RTB Data, Geo Data • Uses all 50 granular data points + Log Level Data • Creates a bidding algorithm that learns over time • To find the best performing ad placements 51
Slide 52: DV's Granular Data Becomes More Valuable - Contextual
2 DV’s Granular Data Becomes More Valuable - Contextual Data Segment Algorithmic Automation Binary Decision, Fixed $3 CPM, Reduced Scale, Models take granular data (multi-segments) to optimize KPI is not considered cost & scale & KPI to choose the best ad placements Universe of Trucks & Truck Access
Slide 53: Improves Workflow Automation
3 Improves Workflow Automation Before: A Trader has to go into their settings inside a DSP and manually make changes to optimize their campaigns DV Data DV Data DV Data Schedules Ad Campaigns Chooses Data Sets Chooses KPI Optimization Advertiser/ Trader DSP Buys Ad Agency Inventory Custom Bidding Mo
Slide 54: Strong Team Drive Value Today & Tomorrow
4 Strong Team Drive Value Today & Tomorrow Remi Lemonnier Co-Founder & President This transaction is critical PHD Machine Learning to help our clients Navigate a future set Julien Hirth Co-Founder & GM of opportunities Engineering and Math That are just around the corner 72 People with 20 DataPlScie
Slide 55: Maximizing Outcomes, Minimizing Risk
Maximizing Outcomes, Minimizing Risk: AI-powered Programmatic Decisioning Joshua Nafman, Vice President, Data & Operations, Diageo 55
Slide 56: BREAK
IDV | INNOVATION DAY BREAK
Slide 57: Message from DV's Customers
Message from DV’s Customers Gian LaVecchia, Managing Director Americas 57
Slide 58: AI @ DV: R&D Scale, Impact and Trust
AI @ DV: R&D Scale, Impact and Trust Nisim Tal, Chief Technology Officer 58
Slide 59: Our Journey With AI
Impactful and Trustworthy AI Our Journey With AI We Build AI, We Use AI, We Sell AI, We Do It Right. AI as a Growth and Value Creation Responsible AI Scale Catalyst With AI 59
Slide 60: AI as a Growth and Scale Catalyst
Impactful and Trustworthy AI AI as a Growth and Scale Catalyst Key AI Investment Areas Customizable-AI to drive 1 Programmatic Optimization (Scibids) Predictive AI Multimodal, Language Agnostic, 2 Content Classification Platform Classification AI Traffic Analysis System 3 Fraud Detection AI (Fraud,
Slide 61: Value Creation with AI
Impactful and Trustworthy AI Value Creation with AI Anomaly Detection Fraud Detection 25% Less humans are needed for 40% At Scale preventative monitoring both Of the new fraud schemes for security and for production. are detected using DV’s AI Use AI to build AI Teams Productivity 66% Savings in lab
Slide 62: With Great Power Comes Great Responsibility
1 Impactful and Trustworthy AI 5 2 Large Scale With Great Power Comes Data Assets, Great Responsibility Technology, and Integrations Does Faster Mean Better? 4 3 62
Slide 63: Capturing the AI Opportunity
Impactful and Trustworthy AI Capturing the AI Opportunity HOW WE ARE SET TO SUCCEED 1 Domain Expertise 2 Proprietary Data and Access to Talent Sets at Scale Israel, France, Finland, Fraud Activity Belgium, Germany Attention Signals Impactful And Computer Vision, Neural Suitability Data Trustworthy A
Slide 64: AI @ DV: Classification Innovation
AI @ DV: Classification Innovation Jack Smith, Chief Product Officer 64
Slide 65: Video Classification Is Essential
Video Classification Is Essential >80% ~10% of all consumer of brand suitability internet traffic was incidents are video traffic in 2022 detectable using Source: Cisco metadata only 65
Slide 66: Video Classification: Completing The Picture
Sentiment Video Classification: Completing The Picture With our Advanced Scene Analysis, Avoidance computer vision, audio and text analysis Categories are combined to provide a complete picture of the content to avoid and the right content to target Contextual Categories Allows Global, Cost-effectiv
Slide 67: Demystifying The Magic
Yan\ Demystifying The Magic ID\V | INNOVATION DAY
Slide 68: Advanced Video Scene Analysis Saves Time and Expense
Advanced Video Scene Analysis Saves Time and Expense Join video, audio and Deep analysis Scan all video Define and textual analysis to of important frames describe scenes determine avoidance, frames context and sentiment Advanced Scene Analysis drives a 20x cost savings over analyzing each frame wit
Slide 69: Looks Like Magic
Looks Like Magic Colgate combined brand suitability, fraud and viewability lower their block rate reinvest savings in higher quality media. >77% The Results Are Real Video qCPM 69
Slide 70: Protecting Media Quality on an AI-Powered Internet
Protecting Media Quality on an AI- Powered Internet Deva Bronson, EVP Global Head of Brand Assurance, Dentsu Media Bill Tucker, Group EVP of Data, Tech, Measurement, Addressable Media at the ANA 70
Slide 71: AI @ DV: Financial Update
AI @ DV: Financial Update Nicola Allais, CFO 71
Slide 72: Financial Takeaways
Financial Takeaways 1 2 3 EXTERNAL INTERNAL CONSISTENT AI Innovation AI Innovation Execution New Products Enhanced Processes Core Growth Drivers Expanding addressable Delivering value / Updated Q3 & FY 2023 market efficiency at scale guidance 72
Slide 73: DV's AI Innovation Continues to Address Large and Growing Opportunity
DV’s AI Innovation Continues to Address Large and Growing Programmatic Revenue Opportunity Of all open internet $123b >85% digital advertising spending (ex Search) $95b 2023 2025 1. Magna Global estimated Programmatic (display + video) ad spend, June 2023 . 73
Slide 74: Scibids AI Expands DV's Addressable Market
Scibids AI Expands DV’s Addressable Market Global Advertising Spend and DV Scibids Addressable Market Addressable Market Forecast 20251 Forecast 20251 Global programmatic ad spend1 >$320b Global Digital Ex-Search Ad Spend $123b Global programmatic ad spend utilizing customizable bidding ~$21b DV Add
Slide 75: Scibids AI Augments DV's Product Revenue Flywheel
Scibids AI Augments DV’s Product Revenue Flywheel Activation Measurement Programmatic Activation Measurement • Authentic Brand Suitability • Authentic Attention $ AI Optimization • Core Programmatic Products • DV Authentic Ad 3rd Party Data DV Data • Audience data • Attention data • Alternate verifi
Slide 76: Scibids AI Five Year Revenue Projection
Scibids AI Five Year Revenue Projection1 ~$100m 58% CAGR $15 - $17m 2024E 2028E 1. Management’s estimate of Scibids’ revenue contribution 76
Slide 77: DV Activation Evolution: Fast Growth of Programmatic Solutions
DV Activation Evolution Fast Growth of Programmatic Solutions 2015 2018 2023 Launched Launched Launching static segments dynamic segments AI activation DV Static DV Dynamic DV Al Activation Activation Activation 1 DV Product Non-ABS ABS Scibids AI Revenue $133m $152m ~$100m 2Q’23 LTM 2Q’23 LTM FY’28
Slide 78: Significant Growth Opportunities for DV
Significant Growth Opportunities for DV Strategic M&A International Current & Expansion • International Channel New Client Expansion New Product Expansion Growth / • Product/Technology Introduction & Acquisition Extensions Upsell • Product/ Technology Adjacencies ~60% 16% 61% 27% of DV’s Top 700 of
Slide 79: Large Product Cross-Sell and Up-Sell Opportunity
Large Product Cross-Sell and Up-Sell Opportunity 13.2x DV Revenue Multiplier Opportunity 1 Plus Scibids Opportunity of DV’s Top 700 customers are 9.3x ~60% using less than 4 products 6.0x 2.6x 3.6x + 1.8x 1.0x Post Campaign Measurement Authentic Attention Social Measurement Core Programmatic Social
Slide 80: AI is Accelerating DV's Scale Efficiently
AI is Accelerating DV’s Scale Efficiently SPEED ACCURACY Increased accuracy of video- ~3x Faster model development 20-40% classification models CAPACITY1 COSTS More cost effective video >50% Free up of labeler capacity 20x classification AI Capabilities Enhance DV’s Ability to Reinvest and Redeploy
Slide 81: Scibids Expected Near-Term Financial Contribution
Scibids Expected Near-Term Financial Contribution Revenue Current Contribution Estimate 1 Financial Profile 1 Current Revenue Model % of Ad Spend Q3’23 $1m Revenue less Cost of Sales as % of ~80% Q4’23 $3m Revenue Adjusted EBITDA margins Near breakeven FY’24 $15 - $17m 1.Management’s estimate of Sci
Slide 82: Q3 & FY 2023 Guidance Updated for Impact of Scibids Acquisition
Q3 & FY 2023 Guidance Updated for Impact of Scibids Acquisition Q3 2023 Guidance FY 2023 Guidance Revenue Revenue $137m - $141m +24% $562m - $572m +25% Low-end High-end YoY growth Low-end High-end YoY growth at midpoint at midpoint Adjusted EBITDA Adjusted EBITDA $39m - $41m 29% $171m - $179m 31% Lo
Slide 83: DoubleVerify By The Numbers
DoubleVerify By The Numbers A Track Record of Delivering Strong Profitable Growth Revenue1 Adjusted EBITDA1 $ in millions $ in millions YoY Growth Adj. EBITDA margin1 +25% +36% 31% +36% 31% 33% +34% +75% 38% 30% 25% 2008 902 5.5T $43m $95m $0 Year Employees at Media Transactions Net Income in Net Ca
Slide 84: Non-GAAP Financial Measures Reconciliation
Non-GAAP Financial Measures Reconciliation ($mm) FY 2018 FY 2019 FY 2020 FY 2021 FY 2022 Net income $3.2 $23.3 $20.5 $29.3 $43.3 Depreciation and amortization 18.6 21.8 24.6 30.3 34.3 Stock-based compensation 1.4 1.7 20.5 21.9 42.3 Interest expense 3.1 5.2 4.9 1.2 0.9 Income tax (benefit) expense (1
Slide 85: Non-GAAP Financial Measures Reconciliation
Non-GAAP Financial Measures Reconciliation a) M&A and restructuring costs for the year ended December 31, 2022 consist of transaction costs, integration and restructuring costs related to the acquisition of OpenSlate. M&A costs for the year ended December 31, 2021 consist of transaction and integrat
Slide 86: Q&A
IDV | INNOVATION DAY Q&A
Slide 87: Key Takeaways
Key Takeaways Mark Zagorski, CEO 87
DoubleVerify Holdings — Investor Day (September 2023) on 6ix