Sylvain Bailly

Instacart not having negative keywords as an option is the best example of what "machine learning" actually costs brands to support. The best solution I've found? Take their auto applied keywords and make their bid the lowest possible (30 cents). I built a really solid ads structure supporting each of our Target sku's and woke up to us bidding on "Ben and Jerry's" as a "optimized auto applied suggesition" for our Vanilla Ice Cream Whey. Gee, I wonder what sophisticated dev team decided "auto bid on product title". Yeah, in the world of CPG there could neeevvverrr be issues with that. Who leads this team and how much do you make. (Hint: too much). Happy Friday to everyone except overpaid leadership that don't actually understand what they're building and get paid boat loads of money for it.

11

4 Comments

Selection bias is at the root of so many bad decisions in retail. Here are three claims I’ve actually heard over the years from either tech vendors pushing their solutions or brand execs blowing their own horn: 1. “Our loyalty program members spend X% more than non-members, so our loyalty program drives higher LTV.” No, it doesn’t. Most likely, what’s happening is that your heavy buyers are enrolling in your loyalty program to get discounts on products they were previously buying at full price. 2. “AOV on our mobile app is X% higher than on our website, so our mobile app drives higher AOV.” No, it doesn’t. Again, what’s probably happening is that the only people motivated enough to download your mobile app are customers who were highly loyal to your brand in the first place. (Also, if you’re running any app-specific promos, see the point above.) 3. “Our post-purchase surveys show that customers come from <insert random channel>, so we’ll double down on our investment there.” I… don’t even know where to begin with this one. When you put it like this, the problem is obvious. Unfortunately, reality is much more complicated than a LinkedIn post, and while these conclusions may never be explicitly stated, they will still live rent-free in the minds of decision-makers if you leave them unchecked. There’s only one reliable, repeatable solution that I know of: Write. Things. Down. Having a system for going from hypothesis to conclusion to initiative is the only way to systematically minimize bias in decision-making. I would love to say “eliminate” instead of “minimize,” but that would be a bit too ambitious, not to mention unnecessary: as in most things, you don’t need to be perfect—you just need to be better than the other guy. #ecommerce #retail #strategy #decisionmaking #data

15

1 Comment

PayPal launching Ad business, leveraging data on millions and millions of shoppers. PayPal, is venturing into the advertising world by leveraging the vast trove of data it has on its millions of customers. The company has hired Mark Grether, who head of Uber's advertising unit, to lead its newly-created PayPal Ads division. The Top 5 Things to Know: 1️⃣ PayPal plans to build an ad sales business around the purchase and spending data it collects from its millions of users, including those on the Venmo app. 2️⃣ The company has already introduced Advanced Offers, an AI-powered ad product that helps merchants target PayPal users with personalized promotions and discounts. 3️⃣ PayPal aims to sell ads not just to its own customers but also to non-endemic advertisers, allowing them to target consumers based on their spending patterns and purchasing behavior. 4️⃣ With data from over 400 million customers and 6.5 billion payments processed in Q1 2024, PayPal believes it can offer advertisers a detailed picture of their target audience. 5️⃣ The move follows a trend of financial companies like JPMorgan Chase building ad networks to monetize their customer data, as they seek new revenue streams beyond their core businesses. PayPal's foray into advertising comes as the company looks to boost growth and find new revenue sources after a post-pandemic slump. By taking advantage of its valuable customer data, PayPal hopes to ride the recent advertising wave. Link to full article from The Wall Street Journal in comments.

22

1 Comment

Ecom Brands: 10 Common Data Issues We Come Across 1/ Attribution Missteps Relying on outdated MTA tools or last/first touch attribution instead of contribution and incrementality models. This can lead to wasted ad spend. Start with simple geolift (MMT) and hold-out tests to measure incrementality. Understand that not all channels can be tracked w/ click-based tools. 2/ Pixel Problems Incorrect, duplicative, or missing ad pixels lead to inaccurate data collection. Audit your pixels monthly for accuracy. GA, Meta, Google, TikTok, Affiliate platforms, etc. 3/ Server-side Tracking Not using server-side tracking, resulting in signal loss and hindering ad optimization. You can be losing 20-30% of your conversions w/o proper SS setup. 4/ Excel Hell & Reporting Inefficiencies Manual and infrequent updating of reports, often trapped in spreadsheets. Automate common reports. Even basic KPIs updated daily for the team will be a big win. 5/ Product Catalogs are Stale Bloated or incorrect product feeds with outdated information and mismatched pixel data. Audit your catalogs frequently. Segment your feeds. Setup alerts. Use a centralized feed source or vendor to streamline updates. 6/ UTM Inconsistencies Poor campaign and UTM naming standards make performance breakdowns difficult. Create a standards document and ensure that all your marketers and agencies use it religiously. 7/ Customer Data Fragmentation Lack of a single source of customer truth, with customer and 1P data spread across various systems. Start to ingest your 1P data into a data warehouse. You should own and know how to govern your customer data. 8/ KPI Blindness Limited understanding of critical metrics like New Customers, LTV, MER, Contribution Margin, CAC... Knowing which to use and when. Understanding driver trees and ensuring your team knows the key metrics for their channel. 9/ Inaccurate Forecast Failures Poorly constructed forecasts or a complete lack of forward-looking projections. Even a basic forecast, shared across the team will help. 10/ Data Silos Isolated data and poor transparency across different teams and departments. Centralize datasets, KPIs and reports where possible. Better communication and data access. The good news is that fixing these issues is 25% technical cleanup and 75% education/process improvement. What other data challenges do you face in your e-commerce operations? #ecommerceanalytics #dtcmarketing #measure #ecommerce

41

5 Comments

The idea of these brand attributes being seen as a 'status symbol' is genius. Here's the ones from Oliver's list that I can really relate to as a CX advocate - and would love my retail brand connections to think about: Password entry, Membership programs, Subscriptions, Resale market, Rental offering... Let your actual customers shop in new ways, let them create value and price accessibility and free up space in their wardrobe or house to buy from you at full price when they need to buy (rather than on discount when you need to sell) No 10% discount pop up... Please. Make this stop. Its not worth it. Flagship stores, Strategic retail partners... Retail grows Ecom - Ecom grows Retail - 1 + 1 = 3. Connect these 2 worlds and allow your customers to pickup where they left off. Remote first, Empty calendars, Well defined rituals, Email is not primary collaboration tool... Talk in open channels, collaborate and progress work efficiently and asynchronously. Separate Conversations (discovery) from Promises (tasks) so both can flourish. Don't confuse being at work (on prem, in a meeting) with doing work.

9

4 Comments

Criteo is rewriting the rules in retail media. 🚀 Their bold moves are reshaping the landscape, and here's why you should pay attention. From prior experience I found their UX lacking, yet their strategic positioning is brilliant. Let's take a closer look... 💡 With over one-third of retail media advertisers managing more than 10 networks, the chaos is real. There's a massive opportunity for a user-friendly solution that simplifies ad buying and ROI tracking as brands look to shift their Marketing Mix drive profit at scale. 💰 Criteo's partnership with Microsoft? Genius. It unlocks fresh revenue opportunities for retailers who want the Ad $ but outsource the tech. (Imagine an ad platform from the creators of Outlook—thanks but no thanks). 📣 By targeting ad agencies that handle non-retail budgets, Criteo could finally solve the puzzle of multi-channel returns. If they nail it, they change the game. I'm betting on it. 💵 With 63% of brands shifting budgets from TV, Search, and Social to retail media, the stakes are high. What are your thoughts on Criteo's game plan? Drop your insights below! [Kudos to Lauren Johnson for the great reporting, as always!] https://lnkd.in/gsvCs5Ph Ready to boost your digital brand? Let's chat: [https://lnkd.in/g8SPGTdH] #retailmedia #ecommerce #advertising

12

5 Comments

According to a new analysis from Lumen Research, Seedtag's DPO (Dynamic Placement Optimization) technology delivers a remarkable 164% increase in attention compared to standard IAB formats 😱 The study highlighted the main benefits of our DPO technology, which are that it: 📈 Maximizes publisher's ad inventory ⚙️ Customizes the ad experience to align with user behaviors 🧑‍💻 Ensures a more enjoyable browsing experience Discover all the findings here: 🔗 https://lnkd.in/ekutZHAF #AttentionMeasurement #ContextualAdvertising #AdTech

18

1 Comment

Am I the only one that gets excited talking about #cpg #marketing? 🤣😂 👏🏽👏🏽👏🏽 Caroline Grace is spot on with her post on pricing and margin expectations, so I had to share. This is where so many #cpgfounders get tripped up…let’s normalize knowing what your pricing strategy is, what retailers expect, and whether or not you can offer what they want and make money — otherwise, you might not be ready. All roads lead back to the 4 Ps… #retailreadiness #cpgindustry #brandmarketing #pricingstrategy #channelstrategy

6

2 Comments

More than just A/B testing, PDQ can work as a meta-tool, A/B testing your tech stack to see what works and what doesn’t. Which upsell tools work best for you? Test it. Which shipping protection do your shoppers like best? Test it. Which post-purchase editing or upsell tool converts best? Test it. We don’t pretend to know things, we just know this one thing: you gotta test it!

24

1 Comment

Ecommerce businesses I would NOT start in 2024: - Single SKU with no recurring revenue component. Too much pressure on first purchase and will crush your MER as you scale - AOV below $35. Too much pressure on CPA. - Anything in the game category. Board games, card games, etc. Too many copycats, Amazon-dependent, and the lowest quality customers in ecomm - Sexual health/wellness. All the restrictions on ad platforms as THC/CBD, without the astronomically high LTVs - Any brand with >10 SKUs that is NOT apparel/accessory. It's a drag on your P&L and overcomplicates your acquisition strategy/forecasting tenfold - Anything seasonal or gifting-driven (flowers, holiday, etc). Makes media buying a nightmare. - Any business where your average gross margin is less than 65% - Any non-vertically integrated "marketplace" style business where you aren't making the product yourself (see above) - Anything liquid. Shipping costs are too much of a drag on gross margin, and the beverage category is a nightmare to stick out in Do you disagree? Am I missing anything? Stay tuned for my next post, ecomm business I WOULD start in 2024 😀

23

9 Comments

Time for part 2... Ecomm businesses that I WOULD start in 2024: - Supplements in underserved categories with massive TAMs and a demo that spends time on Facebook (think: cholesterol, menopause, etc) - Visually driven apparel (streetwear, escapewear, graphic tees, hats) is still the best category to scale on Meta if you have design chops. Low CPMs and unbelievably forgiving of poor website experiences - THC/CBD brands are well worth the compliance headaches thanks to their insane LTVs. You're literally selling drugs. - Anything in telemedicine/telehealth. Go after the long tail prescriptions that Hims & Hers/Ro haven't dominated yet. - Anything with a genuine R&D moat or a new form factor that isn't a liquid in an existing category ("the world's first _____" is compelling for ads, landing pages, and PR) - Anything in the pet category (zero price sensitivity, still growing) - Anything made-to-order or personalized, like astrology jewelery or personally inscribed products (this is a massive acquisition lever) assuming margins aren't terrible - Anything where you have a competitive channel advantage. Are you a paid search/SEO wiz? Launch a line of niche, high-priced chandeliers and dominate the search landscape - Any creator-led business where you're partnering with a celeb with a built-in, relevant audience who understands their role as a content creator for the biz

6

1 Comment

Seamless data access that suits the unique needs of our diverse customers further expanding our robust suite of data access mechanisms. Bringing you “BI Link”

25

New framework: Measurement Methodologies in 2024. Google Analytics and META pixel have real-time data but lack in accuracy -- the numbers look exact but fail in true actionability (ex., signal loss, e.g., people use cross-device, prevent cookies; user behavior change from clicks to zero-clicks [credit: Amanda Natividad], e.g., chatGPT, podcasts, ad impressions). It's easy to stick with click-based analytics (ref. image as Attribution) when you have growth and profitability. The problem occurs when business or marketing pivots are needed; click-tracking lacks the confidence to make big moves (to note: business-specific differences can occur). My vision of growing businesses involves complex analytic capabilities made simple, 'making every marketing touchpoint measurable'. Now, we are getting there as an industry, step by step. I'm putting down my effort to turn this wheel, planned the following for the next year: 1. working with self-service marketing mix modeling (similar approach when Google democratized media buying at a time with Google Adwords prior to the 2010s) [credit: Gabriele Franco] 2. Concepting a CMO-tier cockpit on top of Looker Studio to make Measurement Methodologies simple enough for decision-making 3. Building-up the layer of multi-touch analytics (MTA) on top of Self-reported data 4. Operating with yearly 'Experiment Calendar' for accurate data 5. Pipelined a LLM interface instead of Looker Studio type of dashboards References: * original Measurement Methodologies framework by META Data Science Team, 2024. [credit: Axel Strelow] * enriched by Tommi Salenius, updated MMM, added self-reported, 2024.

35

6 Comments

𝗡𝗲𝘄 – Performance Max for Marketplaces. Hold on, I need to refill my coffee. I look like Ben Affleck smoking a cigarette. Ok. Shake it off. Let's go: 𝗖𝗼𝗻𝘁𝗲𝘅𝘁: Amazon is a product search powerhouse. Polls vary, but a recent single-choice survey showed that 50% of consumers start product searches on Amazon, while a recent multiple-choice survey showed that 66% of consumers start on Amazon. Either way, it's a lot. And that's why having effective organic & paid search on Amazon is vitally important. But for sellers looking to max out their revenue potential, there's another awfully big name in product search: Google. While sellers have been driving Google paid traffic to Amazon and other marketplace listings for years, Google is now making it simple to do so. 𝗞𝗲𝘆 𝗳𝗮𝗰𝘁𝘀: • you don't need a website or even a Merchant Center account • you DO need a Google Ads account • you cannot use an ads manager account since you will be linked to marketplace's ad manager account (along with their MC) • Google doesn't list participating marketplaces; you have to ask your marketplace • conversion tracking is handled by the marketplace – you might want to review how those are measured Linking your Ads account to the marketplace is the most complicated part: three bi-directional links are required. But campaign setup is straightforward. You can run your PMax campaign with just the product feed, or optionally include more assets for additional placements. 𝗔 𝗳𝗲𝘄 𝗸𝗲𝘆 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝗰𝗲𝘀: • no URL expansion • no ACA (automatically-created assets) • no video assets • no cross-account conversion tracking • no NCA mode (new customer acquisition) • no data segments for audience reporting 𝗜𝗺𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗼𝗳 𝗣𝗠𝗮𝘅 𝗳𝗼𝗿 𝗠𝗮𝗿𝗸𝗲𝘁𝗽𝗹𝗮𝗰𝗲𝘀: Amazon sellers that advertise their listings via Google might find this is an incremental audience, meaning they can get a bigger slice of their TAM pie. Some sellers also report that this tactic can boost their organic rankings on Amazon, and that Google has a lower CPA. Still, it's an interesting category of seller that will find this offer most relevant. And one thought occurs to me: this could be a very attractive channel for the Chinese manufacturers that currently comprise >50% of Amazon sellers. Amazon and other marketplaces are already the biggest spenders on Google Shopping & PMax, and this new program will at once decentralize and accelerate that trend. If this program takes off, then over time it could contribute to the CPC wars and race-to-the-bottom pricing that are already squeezing existing Google advertisers. That's why it's more important than ever to differentiate on input & output optimization – what data you're feeding the ad platforms, and how you're measuring them.

183

33 Comments

Running ads for a retail brand with 200K+ SKUs felt like stepping into Amazon’s catalog and being told, “Figure it out.” In DTC, it’s simple. You pick your hero products, build a funnel, and drive TOFU, MOFU, BOFU. Done. Retail? That playbook doesn’t work. It’s messy, layered, and if you treat it like DTC, you’ll burn your ad spend faster than a bad Prime Day deal. Here’s what clicked for me: Retail isn’t just selling products. It’s managing partnerships and prioritizing profitability. I realized I had to think like Amazon does: 1. High-margin products = Sponsored Listings. Amazon doesn’t push everything. It highlights what makes them money. In retail, it’s the same. Vendors offering high margins or paying co-op dollars? Those became my “Sponsored Listings.” They’re worth my ad budget. No margin, no budget? No ads. 2. Market leaders = Best Sellers. Brands like Uppababy or Nuna in baby care are like Apple or Nike on Amazon—people search for them. They’re already in demand. You have to advertise them to capture that intent. 3. Everything else = The forgotten SKUs. Newer brands with no search volume? They’re like that no-name Bluetooth speaker buried on page 37. My approach? Either get the vendor to sponsor ad spend or give me higher margins. If not, no one’s wasting money here. Then, I went deeper. Retail isn’t just about the products—it’s about the people. In baby care, prenatal customers (setting up entire nurseries) were my “Prime Members.” High AOV, high intent. I focused on them: cribs, strollers, dressers. That’s when it clicked. DTC is about funnels. Retail is about leverage. - Leverage your vendors. - Leverage customer intent. - Leverage your margins. If you’re stuck treating retail like DTC, think again. It’s Amazon’s game—only you’re running the entire marketplace.

77

ℹ Google Announces Removal of Sitelinks Search Box On November 21, 2024, Google will discontinue the sitelinks search box feature in search results. This change comes after over a decade of use, with Google noting a decline in usage over time. Key points: ➤ The change will be global, affecting all languages and countries. ➤ It won't impact search rankings or other sitelinks elements. ➤ The Search Console rich results report for sitelinks search box will be removed. ➤Existing sitelinks search box structured data can be left in place without issues. This update aims to simplify search results. While webmasters can remove the related structured data, it's not necessary as it won't cause any problems in Search or trigger errors in Search Console reports. #GoogleUpdate #SEO #DigitalMarketing

4

The Director of Marketing for a $500M merchandise retailer internally reviewed their CDP. Here are 2 things she says the vendor aced and 2 that might get them fired: 🏆//1 Great at Unifying Data She’s faced difficulties integrating past CDPs with the many 3rd-party tools her company uses. It always led to site slow-downs that limited their conversion rates. But her current vendor has seamlessly unified all their customer data across channels. No site slow-downs, no integration issues. And she loves how much it’s helped her better understand customer interactions. She’s also thrilled by how it’s made it easy to craft more relevant, compelling, and effective marketing in real-time for her customers. 👎//2 Not User-Friendly This CDP starts running into trouble with content creation and audience segments. She says, “It’s unnecessarily complex and difficult to manage.” Plus: - Navigating the platform is challenging - And its segmentation tools aren’t as intuitive as they could be - Causing friction in leveraging the CDP’s full range of capabilities 🏆//3 Effective Identity Resolution “We have a high number of customers using different platforms,” she says. So she needs a tool with strong identity resolution (IDR) capacities. The good news: She’s been thrilled with her CDP’s IDR. She says it’s significantly improved their reach “by uncovering an extra 20% of users across different devices!” 👎//4 Lack of Actionable Insights Despite the CDP’s effectiveness, she’s frustrated at how little guidance the vendor offers. “I’ve found myself guiding our CDP more than it’s been helping us,” she says. “It’s not ideal.” She’s hopeful though, as the platform has teased an upgrade that promises enhanced actionable insights. But if the vendor doesn’t deliver — this is going to be a big sticking point for her. She’s felt similarly irritated by a lack of actionable insights in other tools… Like their: - A/B testing AI tool - Personalization tool - Product recommendation tool She doesn’t want to have to waste hours digging for actionable insights. She wants platforms that make it easy for her to get the main points and quickly act on them to drive market growth. ___ Tl;dr A few *amazing* features won’t make up for others that constantly frustrate your clients. MarTech is a competitive market – if you don’t offer the full enchilada, buyers will find a competitor who will. Thoughts?

9

3 Comments

All the rage is "Retail Media"...everyone talking about how best to leverage retail media with advertising budgets. Two alarming statistics. #1 Retail Media will attract more incremental advertising spend than Alphabet and Meta combined. The more alarming statistic? Amazon is forecasted to earn 76% of that retail media advertising spend. Next closest within retail media is Walmart with 6%. Google and Meta have been called duopolies but the "toll booth of eCommerce" has emerged. Amazon keeps collecting and reports this Thursday. https://lnkd.in/gBWUwZSB $AMZN $META $GOOG $WMT

17

1 Comment