Lenny's PodcastTimothy Davis: How Shopify validates new ad platforms fast
Through signs of life tests on 1% Meta lookalikes, Shopify proves a paid channel; agencies kickstart the work, then internal teams own the playbook fast.
CHAPTERS
- 0:00 – 2:06
Why “paid is for everyone”: setting the stage for performance marketing
Lenny and Timothy open with a provocative claim: every company should consider paid growth because platform real estate is increasingly pay-to-play. Timothy frames where paid fits alongside other growth engines and why this conversation matters now.
- •Paid growth is increasingly unavoidable on major platforms (Google, Meta)
- •Organic visibility is shrinking; promoted distribution is rising
- •Paid can complement other engines (SEO, word of mouth, sales) rather than replace them
- •Preview of what “good” performance marketing looks like in practice
- 2:06 – 3:45
What counts as performance marketing (and what “paid” actually includes)
Timothy clarifies terminology: performance marketing is measurable marketing, usually online, but can include offline and affiliates depending on context. The key is aligning on definitions so teams aren’t talking past each other.
- •Performance marketing vs. paid growth: often interchangeable, but clarify scope
- •Online channels are the default, but offline and affiliates can be “performant” too
- •The core requirement: measurement and accountability
- •Why specificity matters when planning budgets and teams
- 3:45 – 6:32
Sponsor break (BuildBetter & OneSchema)
A brief sponsor segment featuring BuildBetter.ai and OneSchema, focused on AI-assisted product insights and CSV/file-feed integrations.
- •BuildBetter: turning unstructured customer data into structured insights + workflows
- •OneSchema File Feeds: building CSV-based integrations quickly without engineering
- •Emphasis on reliability, validation, and operational efficiency
- 6:32 – 8:41
Paid search as the baseline: user-driven vs. disruptive media
Timothy argues that while not every disruptive channel fits every business stage, paid search is nearly universal because intent is explicit. He contrasts search (pulled by user intent) with social/video (pushed into feeds).
- •Paid search is “user-driven” and often the safest starting point
- •Social/video ads are disruptive media and may not fit early-stage needs
- •Even influencer marketing is still a form of paid distribution
- •Practical ordering: start with Google Search before expanding outward
- 8:41 – 9:45
Finding growth potential: where your users already are + “signs of life” tests
Timothy outlines how to identify channels worth scaling by inspecting where organic traffic already originates, then validating with small-budget experiments. He introduces the “signs of life” concept to reduce risk before scaling.
- •Use analytics to find channels already sending traffic/conversions
- •Turn “existing signal” up by adding paid investment where users already exist
- •Run small experiments to confirm channel viability (“signs of life”)
- •Scale only after validating creative, messaging, and funnel readiness
- 9:45 – 12:22
Case studies: Hairstory & Ipsy + why creative doesn’t transfer 1:1 across platforms
Timothy shares examples where companies discovered unexpected platform traction (Meta/TikTok) and built tests around it. A key lesson: creative that works on one platform can fail on another due to different user contexts and norms.
- •Hairstory/Ipsy: spotting platform-driven traffic and testing paid expansion
- •TikTok’s early days required experimentation, not best-practice certainty
- •Platform UX and user mindset change what “good” creative looks like
- •Use outcomes to iterate, not to assume channel incompatibility too early
- 12:22 – 18:56
Running platform experiments well: first-party data, lookalikes, and test design
Timothy gets tactical about early experimentation: start with your own customer data and build tight lookalikes to maximize early signal. He discusses diagnosing failures (targeting vs. creative vs. messaging) and the realities of budget/statistical constraints.
- •Start with first-party customers; build lookalikes (e.g., 1% → broader ranges)
- •Use tight targeting first when budget is limited
- •Diagnose issues via engagement metrics (CTR), then isolate variables with tests
- •Statistical relevance is ideal, but budget often sets test duration
- •Create a learning culture—failed tests still generate valuable insight
- 18:56 – 20:19
Choosing the right platforms: Google ecosystem, Meta, TikTok—and why video is rising
Timothy shares his default platform stack and explains what “Google” actually means (Search, YouTube, Display). He emphasizes video as a major opportunity—if you have a creative production flywheel to sustain refreshes.
- •Google includes Search, YouTube, and GDN—each with different roles
- •Recommended starting order: Google Search → Meta → YouTube (if video-ready)
- •Video can perform extremely well when measured appropriately
- •Creative refresh is mandatory; one winning asset isn’t enough
- •YouTube creative tip: start with emotion (comedy, happiness, resonance)
- 20:19 – 27:34
LinkedIn for B2B: expensive clicks, powerful targeting, and an enterprise playbook
LinkedIn’s costs are higher, but its targeting can be uniquely valuable for enterprise sales and high-LTV products. Timothy shares a concrete story of targeting decision-makers (and their teams) to influence a major deal.
- •Expect LinkedIn CPM/CPC to look ~3x more expensive than other channels
- •Unique power: job titles, industries, and company-level targeting
- •Deal influence example: targeting Coca-Cola stakeholders with specific objections
- •When to use LinkedIn: typically after Google/Meta unless you’re enterprise-first
- 27:34 – 33:31
When to start investing in paid growth: speed, demand, and product-market fit
Timothy explains how timing depends on goals (fast results vs. slow-burn SEO) and whether demand already exists. He warns against forcing search spend for awareness and emphasizes ensuring operational/product readiness before scaling paid.
- •Paid is faster than SEO, but only if there’s demand or a clear strategy
- •Don’t use search to manufacture awareness—use display/video for that job
- •“Coattail riding”: bidding on competitor terms can work for similar products
- •PMF isn’t just desire—it includes operational readiness (e.g., currency/payment fit)
- •Avoid scaling ads when conversion is impossible; you’ll burn trust and annoy users
- 33:31 – 40:37
What agencies often get wrong—and how to work with them without getting burned
Timothy describes why performance often improves after bringing work in-house: many agencies rely on copy-paste playbooks and don’t go deep due to client load. He shares a pragmatic approach: use agencies to start, but plan milestones for in-house ownership.
- •Common agency failure: standardized playbooks and shallow account management
- •High-impact work is often in the weeds: landing pages, query reports, structure, etc.
- •Build an ops cadence (weekly/biweekly/monthly checks) to prevent drift
- •Use forward milestones: start with agency/consultant, then hire in-house at scale
- •Red flag: an agency resisting a transition plan to internal ownership
- 40:37 – 42:50
Hiring for performance marketing: “signal not noise” + first 3 hires
Timothy prioritizes analytical thinking over channel-specific experience because platforms can be taught, but judgment with noisy data is harder. He outlines early hiring: a data-driven growth marketer, then creative support, then a dedicated data scientist.
- •Hire for decision-making with data: identify signal vs. noise
- •Interview tactic: overwhelm with metrics, see if candidates anchor to goals
- •First hire title: Growth Marketing Specialist/Manager (generalist early on)
- •First three hires: performance marketer → creative/brand designer → data scientist
- •Ad copy should be collaborative and relentlessly tested (surprising winners happen)
- 42:50 – 47:34
Sponsor break (Eppo) + why experimentation velocity matters
A sponsor segment on Eppo’s experimentation and feature management platform, focused on faster, more rigorous A/B testing and analysis workflows for growth teams.
- •Eppo supports A/B testing + feature management for multiple growth use cases
- •Emphasis on advanced statistics to shorten experiment timelines
- •Self-serve UI for analysis and sharing insights across teams
- 47:34 – 53:40
Metrics that matter: ad strength, quality score signals, and custom reporting
Timothy walks through how he evaluates Google Ads performance using in-platform diagnostics like expected CTR, landing page experience, and ad relevance. He shows how simple actions—like pausing one bad ad—can materially shift account performance.
- •Use Google’s own diagnostics to find the “hole in the ship” (weakest lever)
- •Ad relevance, expected CTR, and landing page experience guide prioritization
- •Ad strength distribution (excellent/good/poor) can reveal quick wins
- •Quality score and CPC anomalies are investigation triggers
- •Don’t just watch top-line metrics—use platform-level signals to drive actions
- 53:40 – 1:02:12
Competitor analysis: benchmarks via reps, impression vs. click share, and “true competition”
Timothy explains how to benchmark CPC/CTR without generic internet averages—use platform reps and competitor sets. He then reframes competitive visibility away from “ego marketing” (top of page) toward serving the right users, measured via click share, and introduces a “true competition” metric framework.
- •Benchmarks vary by industry; use platform partners for anonymized comparisons
- •Impression share can encourage ego marketing; click share reflects relevance to users
- •Visualizing impression vs. click share helps identify meaningful improvements
- •“True competition” metrics (via Auction Insights) reveal real threats over time
- •Competitor conquesting may be accidental (close variants); validate before reacting
- 1:02:12 – 1:08:52
Attribution and incrementality: multi-touch, conversion lift tests, and when it’s worth it
Timothy favors multi-touch attribution (often time-decay) but stresses attribution alone can’t prove incrementality. He explains GeoX and conversion lift testing, including holdouts, opportunity cost, and minimum spend thresholds to make results meaningful.
- •Preferred attribution: multi-touch with time-decay (users forget first touch quickly)
- •Attribution is biased; it doesn’t answer “would they convert anyway?”
- •Incrementality approaches: Geo experiments (GeoX) and conversion lift tests
- •Platforms can help run lift tests, but you need sufficient spend/signal
- •Rule of thumb: incrementality testing isn’t worth it at very low spend levels
- 1:08:52 – 1:24:41
Building and scaling the team: ops cadence, workload math, and onboarding to impact faster
Timothy details operational practices that keep teams effective: an ops cadence spreadsheet, “hands on keyboards” execution, and a workload calculator to decide hiring vs. cutting meetings. He also shares an onboarding philosophy aimed at reducing time-to-impact from ~90 days to 30–45 days.
- •Ops cadence: defined weekly/biweekly/monthly account actions with accountability
- •Protect execution time—avoid meeting overload and keep operators in the tools
- •Workload calculator: compare quarterly capacity vs. commitments to justify hiring
- •Onboarding: assign real ownership early, set clear expectations, coach via live account reviews
- •Aim to shorten time-to-impact by making “how we do it here” explicit
- 1:24:41 – 1:37:16
Modern realities: ATT/SKAN, creative as a growth lever, and where AI fits (and risks)
In rapid-fire Q&A, Timothy covers ATT’s impact (manageable with SKAN for iOS measurement), argues creative is massively underappreciated, and explains how AI has long been embedded in ad platforms via bidding and recommendations. He also flags potential malicious automation patterns.
- •ATT impact: manageable if you have the right measurement approach (e.g., SKAN)
- •Creative is often the biggest lever—memorable storytelling drives performance
- •AI isn’t new in ads: Smart Bidding and recommendations have used it for years
- •AI is emerging in creative generation/iteration (dynamic builders, asset variations)
- •Potential downside: automated “drip” tactics to slip policy enforcement at scale
- 1:37:16 – 1:42:03
Lightning round: books, media, products, mottos, and career influence
Timothy shares personal recommendations and influences, from stoic philosophy and deep work to favorite shows and a caffeine-reduction product. He closes by highlighting Casey Winters’ impact on his career and an unexpected tennis fact.
- •Book recs: Daily Stoic, Great by Choice, Deep Work
- •Favorite media: X-Men ’97, RRR, The Playlist, Welcome to Wrexham, Billion Dollar Code
- •Favorite product: Magic Mind for focus (placebo or not)
- •Life mottos about expectations shaping happiness and perception
- •Career influence: Casey Winters (plus Casey’s top-tier tennis background)
