Michael Harty

Budget ← [ EXPERIMENTATION TALENT GAP ] → Need There’s a HUGE gap in experimentation & optimization adoption → On one side you have your FAANG + another 50 or so organizations Running 100s to 10s of 1000s of experiments each year. Employing an army of analysts, researchers, engineers, designers, managers & economists. How do you compete with businesses who are so large they need economists to better understand the impact they are having on their users & customers?  (but that’s a whole different topic for another day) → On the other side you have small marketing teams (1-5 people) Barely trying to grasp the complexity of running experiments (let alone at scale) Fighting tooth and nail for budget & headcount Trying to fill the missing expertise with UNICORN 🦄 talent. → Right smack in the middle are all the rest They can’t compete with the giants So they end up levitating towards the bottom of the scale On the hunt for Unicorns as well. Individual employees who can: -Identify (research, analysis)  -Execute (engineering, design, QA) -Deliver (present, document, brief, educate) If you are thinking… “I know there are no Unicorns but I still need to find one or something close to it” You are not alone. And it shows. This week I’m tackling this gap. My goal(s) for this week: - Why it’s dangerous to hunt for experimentation unicorns - What you can do today to try and fill this talent gap - How to compete and win even against giants Happy Monday & have a wonderful week! — — — Still reading… 1. Follow me Hristian 2. DONATE SOCIAL DOPAMINE to a worthy cause: EXPERIMENTATION* *→→ like, comment or share — — — P.S. “Experimentation Confidential | Off the Record” ↳ Private conversations. Actionable insights. → Let’s talk → #experimentation #gtm #b2bmarketing #cro #hiring #employees

5

Bing has bottomed... Well, in advertising revenue, anyway 😂 What does this mean for search advertising as a whole and SME's? Assuming this is just a market recovery after the 2021 boom cooled down, the trendline could continue and potentially knock out a future high as Bing gained market share in 2024, having first broken the 10.51% market share this year. There are underpriced opportunities at Bing, but it's never really front of mind for business owners and marketing teams. While Bing grows its share and before every business ends up spending there, it's worth taking it seriously before the unit economics become too tight. It's also super easy to import your campaigns from Google into Bing with a click of a button!

10

🍪 Google abandons third-party cookie depreciation plans... "After two months of testing, adtech firm Criteo determined that Google’s Privacy Sandbox, the company’s suite of proposed cookie alternatives, would erase publishers’ ad revenues by 60%." Third-party cookies have long been the backbone of targeted advertising, allowing companies like Google to track user behaviour across multiple websites and serve personalised ads. However, concerns over data privacy and online tracking have led to increased pressure on companies to find alternative solutions. However, Google will continue developing Privacy Sandbox APIs and will also debut new privacy measures. For instance, the company plans to offer IP address obfuscation in Chrome’s Incognito mode. Was this inevitable? #thirdpartycookies #digitaladvertising #PPC

3

🔍 New Article Alert: Setting Priors in Bayesian Marketing Mix Modeling (MMM) Bayesian MMM leverages prior distributions to embed domain knowledge, enhancing model robustness amidst noisy or uncertain data. In my latest Medium article, I delve into the importance of setting the right priors to expedite model convergence and yield actionable insights. 📌 What You'll Learn: - The role of priors in reflecting pre-existing beliefs about channel effectiveness, price elasticity, and more. - Practical steps to set priors for better budget optimization and stakeholder alignment. - How to choose between informative and uninformative priors based on available data. - Common distributions used in Bayesian modeling and their applications. Read the full article here to explore how integrating expert judgment through priors can transform your marketing analytics. #Bayesian #MarketingMixModeling #MMM #DataScience #Analytics #MarketingStrategy

8

1 Comment

AdAttributionKit: Unpacking Apple’s new ad attribution framework AAK unifies Apple’s various advertising attribution frameworks under a single umbrella product name and extends SKAdNetwork in meaningful ways, including with support for re-engagement and alternative app marketplaces. These are substantive improvements. But impediments to adoption remain: IP-based attribution is still widely viewed as a suitable alternative to Apple’s native iOS attribution framework, meaning that platform tools like Conversion APIs will take adoption priority.

101

5 Comments

Huge news from Google overnight. However, from a Content Ignite perspective it's confirmation that our strategy to double down on ID solutions is the correct one. Everything suggests Chrome will end up with a serious erosion of 3PC, it's just going to be a slower burn than the cliff of total 3PC cookie deprecation. Our new ID partnerships with Prebid, Live Ramp, The Trade Desk, EUID, UID, ID5, Panorama (Lotame), Audigent and more will be rolled out across our publishers, this is exactly the type of scenario we built our Fusion platform for. The ability to split test these solutions with ease, feedback the data to our customers, and optimise represents a significant uplift opportunity for our publishers. More importantly, I'm looking forward to no more cookie memes/puns...

36

📊 The New AppsFlyer Performance Index is Out — Here Are My Key Takeaways 👇 🚀 The latest Performance Index from AppsFlyer is live, and it’s packed with insights on which networks are driving serious app growth. But instead of diving into every detail, let me highlight a few of my favorites that stood out in this edition 💡 Top Channels (for iOS): 1️⃣ Apple Search Ads: Still the #1 player for iOS growth, delivering high-quality users, especially as privacy changes reshape the mobile landscape. If you’re not fully leveraging ASA, you’re missing out on serious potential for scaling iOS user acquisition. 2️⃣ AppLovin: A dominant force in mobile marketing, known for its ability to deliver impressive results across both gaming and non-gaming apps. 3️⃣ MOLOCO: Emerging as a top contender with strong performance in programmatic advertising, especially for scaling apps across regions. 4️⃣ Meta Ads: No surprise here. Meta continues to offer massive reach across platforms, remaining a key channel for both iOS and Android. 5️⃣ Google Ads: Another powerhouse that combines scale with performance, delivering great results for cross-platform growth. While these networks continue to dominate, the key isn’t just knowing the rankings — it’s about how you use this data to fuel your strategy. Whether you’re optimizing for quality users on ASA or scaling across broader platforms like Google Ads, having a smart growth approach is everything 🎯 And if you need guidance on how to optimize Apple Search Ads (or any other network) for maximum impact, SplitMetrics is here to assist 💪 #AppGrowth #MobileMarketing #AppleSearchAds #PerformanceMarketing

60

3 Comments

📌 From Baby Steps to Giant Leaps: How Seraphine Leveraged Pinterest Ads for Explosive Growth Using last click attribution, social media and Pinterest in particular showed little value for Seraphine. Seraphine intuitively knew the real issue and challenge was one of measurement and looking through the right lens. Testing with Fospha and moving to a full funnel measurement strategy Seraphine uncovered significant opportunities for paid social and Pinterest as a result; 📌 Achieved a 13X increase in channel revenue in the UK market 📌 Improved CPA by 50% and ROAS by 96% in the first half of 2024 compared to the same period in 2023 📌 Increased Pinterest channel spend by 590% year-over-year Read the full story here on Fospha's Website https://lnkd.in/eFZh24kQ

21

Are you losing 30-40% of your Facebook ad attribution data? That's what we're seeing across ecommerce brands who haven't implemented Meta's Conversions API. But some marketing teams struggle to justify investment to leadership. So I spent one afternoon with #Ai and built a simple tool... Introducing our Meta CAPI ROI Calculator - A free tool that shows you exactly what you're leaving on the table. Just input your: • Monthly Meta ad spend • Average order value • Current conversion rate • Average cost of customer acquisition And discover: • Expected CPA improvement • Net attribution improvement • Potential incremental revenue from going server side One luxury retailer we worked with discovered they were missing millions in attributable revenue annually. They implemented CAPI and saw a 121% increase in reported conversions within 10 days. Ready to see your numbers? See the comments for access to the calculator. P.S. As server-side tracking specialists, we're here to help if you decide to make the move. Let's chat if you're ready to show your boss marketing results and KPI improvements. #digitalmarketing #ecommerce #metaads #serverside #roi #analytics #retailers #innovation

7

8 Comments

One of the reasons that we started pubX was to help quality publishers to thrive in a world that felt stacked against them. A driving philosophy behind this was that Platforms, such as (but not limited to) Meta, were able to skirt the same editorial standards that (particularly news) publishers had to adhere to, especially regarding the content shared on their platforms, and through their resultant scale, hoover up ad spend. With this latest announcement, Meta has renewed their commitment to shirking responsibility, and taken a backwards step. This is not only a negative for publishers, but has a real impact on society. It is our hope that advertisers will start choosing quality over convenience - it won't just be Publishers that benefit.

11

From comparing David Beckham to Google Analytics back in 2012, to jumping on the very demure / Brat summer trends, I've always loved spuriously linking topical trends with data and optimisation. NewDigitalAge were kind enough to publish my most recent musings. With various challenges that brands face in getting people to the site in the first place, making the most of each visit has never been more important. From understanding the right context and associated message, to good old-fashioned smooth user journeys, there's so much to gain by maximising the POST CLICK EXPERIENCE (PCE) https://lnkd.in/eqKANBKu #postclickexperience #ux #conversionrateoptimisation

10

5 Comments

With significant ad spend still disappearing into the 'unknown delta', it's time to ask: Who benefits from keeping the programmatic advertising ecosystem murky? In the complex labyrinth of advertising, transparency frequently takes a backseat. Recent news stories have shown that even well-known brands and agencies can support questionable ad networks without meaning to. This has to do with tracking and spending every dollar, not just where the ads end up. Why does this opacity persist in an era when clear ethical standards are expected? Not only is there an intricate web of programmatic advertising, but there are also entrenched interests that resist change. We need to push for a change in direction toward more clarity. Transparency should not be an added bonus; it should be at the very core of all of our operations. Only then can we build trust and meaningful innovation in advertising. Now let's talk about how we can encourage a more transparent AdTech environment. Image: Maciej Zawadzinski

7

Over the past 25 years, advertisers have had to pay incredible amounts of money to reach customers online. Why? Advertising auctions. To buy advertising slots from companies like Google, Facebook, or Amazon, your favorite brands constantly participate in auctions against each other. Big Tech companies usually allow the highest bidders to display their ads. That competitive environment forces brands to outbid each other to win ad spots - making them pay more and more. ‍ It’s taken decades for advertising costs to rise so much, but today, when your favorite brand sells a product for £100, they probably had to pay £20-£40 on ads for Big Tech to find the customer to buy that product. That’s a whopping 20% to 40% of an item’s purchase price on ads just to get one single sale! ‍ Big Tech has gotten richer, but shoppers – the people advertisers are trying to reach – haven’t seen any of that. Change is coming.. https://lnkd.in/eriGaG8u

7

While digital functions like PPC and SEO often get funded and resourced without question, CRO is often expected to “prove” its value first. This forces CRO teams to chase quick wins, running surface-level tests for short-term growth. But real value in CRO comes from understanding users deeply—the kind of insights that drive sustained growth. Here’s the problem: CRO can’t deliver long-term results when it’s boxed into showing short-term wins. PPC doesn’t need to justify itself upfront. Neither does SEO. Yet CRO constantly has to prove its worth. This keeps teams focused on tactics, not strategy. Imagine what could change if CRO was given the same trust. More meaningful tests. Deeper insights. Real growth. It’s time to let CRO be a driver of user understanding, not just UX tweaks. #CRO #Experimentation #Growth #UserExperience #Datadriven

22

8 Comments

Most ecoms overutilize Meta & neglect Google hardcore—a huge mistake. Here’s the 3-part, modern playbook top brands run: This isn’t anecdotal… - Based on aggressive ad testing - Across both Meta & Google - For 10+ high-growth brands - All between $5M-$75M in rev Here it is: 1. TOFU PMAX (Prospecting) - Non-brand, new customers only - Single asset group * Later: Test separate PMAX - For high margin products, cats, etc. * Large accounts: Test Demand Gen (needs time) - To feed PMAX high intent, cold traffic - Requires great video; arbitrage from Meta - Test all types: statics, videos, & feed - Run Lookalikes (only campaign that has) 2. MOFU Standard Shopping - For manual bidding - So can control “branded ad” cost - 2nd PMAX gatekeeper to funnel off brand terms * Later: Harvest keywords - To test Generic PPC * Large accounts: Test Display Remarketing - E.g., YouTube - Audience = visitors (non-purchased) 3. BOFU Exact Match Brand PPC - Target Impression Share (95%+) - Should have massive ROAS - 1st PMAX gatekeeper to funnel off brand terms Stop sleeping on Google. Diversify and get all your eggs out of Meta’s basket 👊

27

4 Comments

Following the likes of Robinhood & Salesforce, brands are betting the farm on owned media, but 90% of b2c brands are overestimating their upside. 2 things that differentiate those marquee efforts from most companies: Mass Appeal - Sherwood (Robinhood) and Salesforce+ are going after 2 of the largest audiences, respectively, in consumer investing and sales/marketing. But let's say you operate an insurance marketplace, a telehealth company, or... a mattress company... What are the real odds consumers subscribe to your newsletter? Investment - Sherwood is not messing around. They've hired a cadre of top talent, and then went and bought Chartr to sprinkle on top. You need to be creating a TON of great content to have a chance at creating a real consumer media brand, and too many owned strategies are being built on the back of 2-3 person teams. The truth is... b2c brands need to find audiences offsite, by leveraging their content/data into earned. Tell interesting stories, and create relationships that allow you to syndicate those stories at large.

24

1 Comment

NBD, Chalice AI is included in Ad Age's "5 Biggest Ad Tech Stories of 2024"... ahead of Elon, antitrust, cookie whiplash and MFAs. AI FTW! "AI has made its way into every corner of the ad business: Ad agencies are partnering with generative AI platforms to build up their services; Google, Meta, TikTok and Amazon are creating automated ad products based on AI decision-making; and within ad tech, AI is increasingly being used to power programmatic ad campaigns. Demand-side platforms are developing GPTs—generative pre-trained transformers—which are based on the same large language model technology that powers OpenAI’s ChatGPT. The DSPs are trying to use the GPTs to help ad buyers plan campaigns, and eventually make most of the decisions in programmatic advertising. Some ad tech firms, such as Chalice AI, are experimenting with custom algorithms that brands can use to make smarter ad buys. Meanwhile, Google has expanded its AI campaigns called Performance Max and Demand Gen; Meta continues to push its Advantage+ platform; and Amazon redesigned its DSP this year with AI flourishes, including expanding its Performance+ program." Great roundup, Garett Sloane!

153

https://lnkd.in/dwBSuhU9 This week I look into the key attributes for Managed (Non-Programmatic) Pricing. This video is really just intended for people in Pricing and Yield Teams. In particular I take a pretty strong stance on a few attributes that I see teams trying to incorporate into their pricing decisions.   Spoiler Alert - If you are trying to price Cross Elastic effects and Time-Of-Day effects in a Managed Environment, don’t waste your time!   Controversial Take: You are probably spending too much time on incorporating Sell-Through-Rate and time-of-week or time-of-year into your pricing.

22

https://lnkd.in/eDt_9dZr Interesting and not totally suprising given the can kicking for the last few years. User privacy will act as a conduit for "third party deprecation" (Ask app not to track..) Ad blockers still can wreak havoc on your ad tracking. Identity resolution and attribution away from third party cookies as a strategy should still be important for everyone buying cross channel media.

4