Ad Placement Optimization

The Secret to Finding Your Best Ad Placements (Without Guessing) Most founders run ads like they’re playing darts blindfolded. Targeting? Guess. Placements? Guess. Results? 🤷 Here’s the truth: → 80% of your budget might be going to the wrong placements. → But the right 20%? That’s where the profits are hiding. Let me show you how I helped a DTC brand fix this and triple their ROAS. 👇 It was 11:23 PM. Client pinged me on WhatsApp: “Sales are down. CTR is bad. Should I increase budget?” I said no. Instead, I pulled up their breakdown report. → IG Stories = 90% of clicks → Facebook Feed = 65% of spend → Messenger Ads = 0 conversions The budget was bleeding on dead placements. So we flipped the script: ✅ Cut off Messenger, In-Article, Reels ✅ Doubled IG Stories & FB Feed ✅ Built creatives 100% tailored for those formats ✅ Created a “winning placement loop” (more below) Within 4 days: ✅ +2.7x ROAS ✅ +74% lower CPM ✅ Client said: “Bro... I should’ve done this months ago.” Here's the placement loop we now follow for every campaign: 1. Let Meta auto-optimize for 3-5 days → Don’t restrict, observe. 2. Open the Breakdown Report → Go to: Ad Set > Breakdown > By Delivery > Placement 3. Double down only on proven formats: → Pause what’s not converting. 4. Custom creatives for top 2 placements only → (If IG Stories is #1, design for IG Stories, don’t repurpose.) 5. Re-launch as a "Focused Winning Ad Set" → 80% budget → 20% best placements This is one of the fastest ways to scale profitable ads especially if you don’t have $10K/month to test. Let’s stop the guesswork and make your ads convert like they should. ——— Salman Munir | CEO, AdLinked Helping DTC brands & founders scale Meta ads profitably (no fluff, no gimmicks)

Did you know there are Meta ad settings that automatically tailor your ads to each individual person and most media buyers don’t know how to take advantage? Here’s a full breakdown. 1️⃣ How dynamic placements work Meta is a people-based platform. It first figures out who to reach in an audience. It then decides where it should reach them. It then tailors the ad that is shown to best align to each individual’s preferences. Let’s say Meta decides I should see an ad from a sneaker brand. It then asks: should it reach me on Instagram or Facebook? With a Stories ad or In-Feed ad? With or without a headline? With an image or a video? Even if Meta figures out that I’m the right person for the ad, but it doesn’t show me the ad in the optimal context, I may not click. After all, I only click 1-2% of the ads I see. When you give Meta more options, more breadth of choice, and more variation, it increases the probability that it will find the best person-placement fit. 2️⃣ When to use dynamic placements Given you can benefit from Meta doing this high degree of tuning and tailoring for you, use dynamic placements and dynamic creative as a default. Sure, there are situations where it isn’t the right choice, but that’s the exception versus the rule. Meta has more historic engagement data, behavioral data, and affinity data than anyone else. And it has invested billions into machine learning systems that can turn this data into better ad performance. It knows that I’m more likely to click an Instagram feed ad than anything else. It knows I prefer carousels over videos. It knows I like to Learn More over Shop Now. Use this to your advantage. 3️⃣ Proper Setup Now that you know that these auto-optimizations are occurring, here’s some things to consider. You want to feed this system the right raw materials. Put in a wide variety of inputs. Don’t just turn on auto-placements; upload different crops of your images and videos so that they look great in vertical and square formats. Ad text overlays in case the visual doesn’t show up with the text headline you were expecting. Make your ad adaptive and flexible, so that no matter what version gets shown, it looks great and is hard working. If it’s not clear what you’re going to get, use the Advanced Preview tool so you can see all the placements, all the formats, and all the permutations of your ad. 4️⃣ Summary The secret to outperforming your competitors with Meta Ads is to unlock all the power of its machine learning systems. Understand how they work, and then feed the system all the raw materials it needs to perform optimally. Ensure your ads look great no matter which tailored arrangement is shown to any given user. ————— Interested in this topic? This is one chapter from the Master Class on Meta Optimization Algorithms I’m holding on May 1st. Check the comments for an event registration link.

Google quietly updated its documentation to clarify that each ad location runs its own separate auction. 🤫 This means the top ad slots are decided in a different process than other ad placements on the page. While this may sound like a minor wording change, in practice, it could have big implications for advertisers. ➡️ Your ad could appear multiple times in different positions on the same page. ➡️ The auction for the top spots might be even more competitive, favoring advertisers with a lot of money. ➡️ Lower positions could become more unpredictable, as they’re not directly influenced by the competition for top spots. ➡️ Strategies focused purely on "winning the auction" may need adjusting, because now there are multiple auctions happening at once. Let's look at an example in practice (a very back-to-basics example at that): • You're bidding on a keyword. • You set a max CPC of £5, assuming you’ll compete against all advertisers for any ad position. • In reality, you’re entering separate auctions - one for the top slots, another for the lower positions. • A competitor with a higher bid and better ad rank wins the top auction, but you still win a lower ad placement at a cheaper CPC. • Meanwhile, another advertiser, with an aggressive bid strategy, wins both the top and lower auctions, meaning their brand appears twice on the page. It’s not just about bidding high enough to "win" - it’s about knowing which placement delivers the best ROI. If lower positions convert at a similar rate for less cost, it may be more profitable than overpaying for the top slot. On the flip side, if competitors dominate multiple placements, it could push smaller advertisers out entirely. This does reinforce the need to analyse ad placements, conversion rates, and cost efficiency rather than just blindly chasing top positions. Remember to dig deeper into auction insights and adjust bid strategies accordingly. #ppc #ppcchat #googleads #googleadstips

Amazon dropped one of the biggest Sponsored Products updates we’ve seen in a long time (and personally was waiting for): ->audience targeting inside SP and the ability to create custom ones with AMC. Most advertisers are going to butcher this. They’ll pile audience bid boosts on top of existing ranking structures, placement modifiers, legacy bids — and then wonder why their campaigns nosedive. That’s how you torch your signals and bleed efficiency. Here’s the approach — the same structure we’re running across multiple brands: 1. Build your audiences inside AMC. This is the only place you’ll get truly clean data. Do not be lazy and use the ones you have available by default, they are mid to upper funnel audiences (which might work if that is what you wanna go after). But now you have access to AMC, so no excuse not to customize the audience based on your target. How to: Go to your ad console -> measurement and reports ->AMC → Use Cases → Audiences → pick the behaviour (ATC, PDP views, click-no-purchase, etc.) → create to Audience Hub. 2. Do not slap audience modifiers onto existing campaigns. If a campaign has a purpose — ranking, defence, etc — stacking audiences on top of it just corrupts the whole bidding logic. And if you’re already using placement modifiers, mixing them with audience modifiers is a guaranteed mess. 3. Create a separate SP campaign built only for the audience. Low base bid - start with half of the lowest suggested. Attach the AMC audience. Modifier applies only to that audience (at least 100% and increase this as needed). This isolates the traffic, preserves signal quality, and gives you a clean testing lane. The outcome across every brand using this structure has been identical: higher conversion rate, lower CPC, better margin. Same budget — just higher-quality traffic. The rule is straightforward: pick the audience that aligns with your objective. Don’t target everything. Fix the biggest gap in your funnel first. I’ve mapped out every audience, organized them by funnel stage, and included recommended starting points. Comment ME and share this post, and I’ll send you the file. #amazonad #amazonadvertising

The biggest miss we're seeing with upper funnel execution is FREQUENCY. We just audited a business running awareness media on paid social: - Spending about 16% of budget on awareness (Ad recall optimized) - Running on automatic placements on Meta - 15 unique ads per ad set - Targeting Broad audiences (18+, US, Women) - Average weekly frequency across their awareness efforts was 0.90. We fortunately have a lot of research quantifying how frequency impacts effectiveness (example research shared below). Reach of course still matters a lot, with the first impression having the largest marginal impact. But proper frequency is a key driver of impact: - 0-1.5 weekly frequency is too low - particularly for lesser known brands or more complex products - 1.5 - 2 weekly frequency is optimal - especially when looking at purchase intent - 2+ weekly frequency is most critical for shorter term campaigns, like BTS, but the marginal return per impression is approaching 0 at this point. This is why I have previously harped on broad targeting for "new to awareness" advertisers on Meta. Broad audiences are not inherently bad, and in fact broad reach is important. But broad audiences at the expense of frequency is not a worthwhile tradeoff for most brands moving up funnel. Over emphasis on diverse creative only exacerbates this frequency problem. Performance-oriented teams are conditioned to needing a LOT of ads to "feed the algo" and keep performance up (ASC campaigns often feature 25+ ads). But excessive-creative proliferation dilutes your message and branding even further. So how can you tackle this problem? - Construct your audience size to allow for proper frequency at your budget - broad but not too broad - Test optimizing to reach with a target frequency - this will allow you to maintain a broader audience, but focus delivery on proper frequency - Filter out unproductive reach/impressions with placement exclusions - focus on feed, reels and stories, or just feed and reels. - Limit creative to 2-4 concepts - avoid running with an ASC 20+ creative strategy - Monitor & report on weekly frequency - it's ok to adapt mid flight if you are seeing too high or too low frequency And for those looking to go the extra mile, setup a lift experiment testing across different frequency levels (3-cell, 1, 2 and 3 weekly frequency - or something similar). The above are just guardrails - the best solution for you is to test for yourself. New Engen

Last week, the Marketing Expert Journey channel officially introduced a Google Ads bidding strategy focused on optimizing profit. Unlike the ROAS strategy, where the campaign goal is revenue, the GPOAS (Gross Profit On Ad Spend) strategy focuses on generating profit (sales revenue minus the cost of goods sold). This approach allows for much better ad optimization, especially when the percentage margin varies across products. So far, some advertisers have addressed this issue by segmenting products with different margins into separate campaigns and adjusting the target ROAS in those campaigns accordingly. Unfortunately, this approach has certain limitations: → Splitting campaigns reduces the data volume available for the optimization strategy, leading to slower learning. → The product sold isn’t always the one advertised. A user might click on an ad for a high-margin product but end up purchasing a lower-margin product. A precise optimization signal can be generated solely by using the actual margin from the transaction as the conversion value. Until now, this could be done by modifying the conversion value (preferably via server-side tracking to prevent third parties from accessing margin information). Now, this functionality is being introduced as a native feature in Google Ads. → To use this new feature, conversion reporting with cart data is required, where, in addition to the product price, the product feed also includes COGS (Cost of Goods Sold) for each product (see Google Ads help article: https://lnkd.in/d_vFrWJp). → These strategies are only available in Shopping Ads and Performance Max campaigns. → One of the requirements for accessing this beta is at least 100 conversions per week and spending over $7,000 per week in the campaigns where you intend to use this strategy. During the presentation, Maciek Pikula stressed that maximizing return on investment is unlikely to align with business goals, as a high ROAS usually means a relatively low sales volume. I can add two articles from my end: How to maximize profit from campaigns and how to determine the point of maximum profit: https://lnkd.in/dqz-K8pk Model for maximizing sales with target GPOAS: https://lnkd.in/d635Wh5N

We spend $5.5M/mo on Google Ads. Here's our Demand Gen checklist: 1. Consolidate for Data Density, Segment by Audience. Do not split Ad Groups by placement (YouTube vs. Discover). Demand Gen is designed to be cross-channel; forcing a placement "chokes" the AI’s ability to find the cheapest conversion across the ecosystem. Instead, use Ad Groups to segment by Audience Theme (e.g., Top 25 Search Terms vs. 2% Lookalikes). This keeps your data signals concentrated in one place while allowing you to see which buyer persona is actually biting. 2. Check image AND video asset quantity. Ads & assets > Assets. You need variety in both formats. We've seen 20% more conversions running image and video together vs video only. 3. Check headline and description variety. Campaigns > Demand Gen > Ads > Headlines and descriptions. Demand Gen serves across YouTube, Gmail, and Discover. Each placement renders differently. You need variety. 4. Check view-through conversion tracking. Demand Gen drives view-throughs that Search doesn't. We value them at ~30% of click-through. If you're not tracking them, you're undervaluing the entire campaign. 5. Segment device performance. Campaigns > Demand Gen > Segment > Device. Demand Gen skews mobile because of Gmail and Discover. I've seen mobile CPA run 3x worse than desktop. And other times where mobile significantly outperforms. 6. Review your bid strategy. Settings > Campaign settings > Columns > Bid strategy. For Demand Gen, start target CPA at 2x your Search CPA. CPCs are cheap ($0.09 vs $7+ on Search) but conversion rates are lower. Give the algorithm room to learn. 7. If running product ads, check feed health. Google Merchant Center > Products > Diagnostics. Disapproved products mean your ads don't serve for those items. 8. Set your budget for a "Conversion Floor." While the "15x tCPA" rule is great for rapid scaling, it’s often unrealistic for high-CPA accounts. At a minimum, aim for a daily budget that supports 1–2 conversions per day. If your Target CPA is $200, a $300–$400 daily budget is your "efficiency floor." It will take longer to exit the "Learning Phase" than a $3,000/day budget, but it prevents the algorithm from overspending on low-quality placements while it searches for a baseline. 9. Build custom segments from your top 25 converting Search terms. Audiences > Audience manager > Segments. You're reaching people before they search. That's the whole point of Demand Gen. 10. Confirm auto-tagging is on. Settings > Account settings > Auto-tagging. Without it you lose conversion and view-through data between Ads and Analytics. 11. Retract bad leads within 24 hours. This matters more on Demand Gen because the audience pool is broader. This is 11 out of 31 items we check on every Demand Gen launch. One wrong setting can burn thousands before you catch it. Want a high res copy? Comment ‘Sheet’, send me a connect and Ill DM it to you.

You finally land a newsletter sponsor. You run their sponsorship. It flops… no clicks. No conversions. IMMEDIATELY pull up this post and do these 3 things: First, define what a “flop” is. The first metric I look at is the ad click-through rate (CTR). If your ad CTR is below 0.5%, that’s when you should start to worry. Let’s do the math: • You send to 2,000 readers • 40% open rate → 800 opens • You charge a $30 CPM That means your sponsor paid $30 for that send. At a 0.5% CTR, that’s 4 clicks = $7.50 per click. Compare that to Meta ads, where the cost per click is usually $1–2. So your sponsor just paid 4-8x the price of Meta for less data. No bueno! Now, if you’re in B2B, higher CPCs ($5–10) are fine because the products are high-ticket ($1k-10k+ LTVs). But once you’re above a $10 CPC, you’re typically going to disappoint performance marketers. Here’s what to do to improve ad CTR for next time 1. Sell packages, not one-offs It’s rare for someone to see a single ad and instantly buy. That’s why I recommend selling 3–5 ad packages. Be willing to discount 10-20% for packages. It gives the sponsor multiple touchpoints. You can tweak the creative between issues so that the odds of one version overperforming increase. 2. Diagnose what kind of problem you’re dealing with If every sponsor performs poorly, it’s a systemic content/reader quality issue. If just one flops, it’s a sponsor/ad unit issue. When it’s systemic, look at: • Audience quality (are ALL reader sources low CTR?) • Placement (is the ad buried halfway down the email?) • Content (does the ad clash with your editorial tone? Is the copy feature-driven instead of benefit-driven?) But if it’s a one-off issue, don’t change your whole system; fix that ad. a) Copy is usually the biggest lever Instead of writing in features (“We’re HubSpot, use our AI enrichment”), write in terms of outcomes (“Personalize your outreach and book 20% more sales calls from email”) b) Add proof: “The average email response rate with [our nifty personalization feature] is 30% higher!” c) AB test creative details: • Button vs. no button CTAs • Image vs. text-only ads • Above vs. below the fold primary placements • 3 links instead of one 3. Offer make-goods before they ask If you know the ad underperformed, reach out first. Say something like: “We didn’t see the engagement we expected. Here’s what I think we can do.” You could: • Give them a free rerun in the primary slot • Offer 3 quick-link mentions in future issues • Or remarket to everyone who clicked their first ad • Offer specific changes to the copy/image informed by sponsorships that HAVE worked Half of the B2B sales game is making people feel good and having a point of view. You’d be surprised how often you can win a 2nd chance by coming in with a clear POV on why something underperformed and 2-3 tactical things to improve the 2nd at-bat.

I’ve helped 5 eCom brands exit for ~$500m. The acquirer always wanted lower CPAs: So we pull 8 levers: 1. Creative → Target ~1 new concept per $10k in monthly spend. → At $500k/mo, that's 50 concepts. → 70% video (top of funnel, builds awareness) → 30% static (bottom of funnel, closes sales) That's 35 video concepts, 15 static concepts. Then 2-3 hook variations per video, and 5-8 variations per static. That's roughly 70 videos and 90 statics. Cut 70%+ of creatives before they hit two weeks. Your top 1-2% of ads should drive ~50% of spend. In most accounts, 70-80% of creative continues performing month-over-month. That means: → To maintain: replace 20-30% monthly → To grow 20%: replace churn + add 20% more volume 2. Media buying There are three actions that cut CPA without new ads: → Pause or spend-cap everything above target CPA → Retest old winners with new copy, headlines, landing pages → Scale the top 1-2% to take ~50% of total spend 8-figure brands can cut CPAs by 50% with media buying alone. Keep testing budget under 20% of total ad spend. Limit budget changes to 10-15% max, but make changes twice as often. 3. Website optimization The benchmarks: → CVR: 3%+ (top 10% hit 4.7%+) → Add-to-cart: 7-10% → Checkout completion: 60%+ Sometimes a landing page with 10% higher CPA leads to faster repurchases and higher LTV. 4. Subscription optimization The targets: → Monthly subscription churn: under 7% → 12-month retention: 40%+ → Repeat purchase rate: 30%+ The lever is segmentation: → Subscription vs one-time buyers → 4 week vs 8 week vs 12 week frequencies → Product categories → Acquisition channels The gap between 2x and 4x purchase frequency is a 2x LTV multiplier. 5. CRO Target email opt-in: 2-5%. Run distinct landing pages for each avatar. Example avatars for a supplement brand: → General nutrition → Gut health → Weight loss 6. Tracking optimization Click-based attribution overvalues lower-funnel performance by up to 250%. Top-of-funnel creative can drive 13X more incremental acquisitions than bottom-of-funnel. Click attribution will tell you the opposite. Post-purchase surveys catch what click attribution misses. Track individual nCAC on every ad you run. 7. Ad copy and headlines Ad copy can boost performance by 30%. Give creators selling points, not exact scripts. Target: → 40%+ hook rate → 2%+ CTR → 2-3 hook variations per video concept minimum 8. Data reporting and analysis Know two numbers: Maximum spend (company stays profitable): → Gross margin - OpEx = maximum marketing spend % → Example: 50% margin - 10% OpEx = 40% max Target spend (customer stays profitable): → Project 3-month customer profitability = your target CPA → Example: $55 AOV, $30 first purchase profit, $39 at month 3 = $39 target CPA End of the day, acquirers want: → Profitable customer acquisition → Reliable new customer growth for 3+ years → LTV and margins optimized

OOH - not your grandmother's ad medium anymore. Too many brands still treat OOH like a legacy channel. The biggest mistake marketers make with out-of-home? Treating it like a one-off creative play instead of a performance channel that can be planned, measured, and optimized with data. Here’s how to plan OOH with the same data-driven mindset as digital: 1. Define Your Audience—Beyond Demographics Use MAIDs and purchase data to map where your actual customers move. Combine first-party CRM data with behavioral insights for smarter targeting. → Example: A B2B SaaS brand skips generic business districts and targets high-value commuters from key zip codes. 2. Pick the Right Markets & Placements TRP (Target Rating Points) = reach & frequency in a given city. Impression quality > quantity—pedestrian vs. vehicular traffic, dwell time, viewability. Competitive Share of Voice—are you cutting through the noise? → Example: A CPG brand chooses high-frequency transit ads in NYC but large billboards in LA. 3. Measure Performance—It’s Not Just Awareness Geo-lift & MAID exposure: Track real conversions post-exposure. Branded search lift: A strong proxy for OOH impact. Brand lift studies: Pre/post awareness shifts. → Example: A fintech brand saw a 47% lift in branded search and 22% more app installs in their OOH markets. 4. Optimize & Scale Double down on best-performing placements. Test CTA-driven vs. brand-led creatives. Shift spend dynamically across billboards, transit, and place-based formats. → Example: A DTC brand found subway ads outperformed street-level placements and reallocated 40% of their budget. OOH isn’t old-school—it’s a measurable, performance-driven channel when done right. Start with data, measure everything, and optimize.