Tips for Improving Ad Set Performance After Learning Phase

Stop overthinking when your FB ROAS is below 1.2x. I've found that I only need to focus on 4 things (step-by-step). It's stressful when performance is poor. I've been there. I've made many mistakes and learned from them. Here's how I do it now. PROCESS: 1. 𝗧𝗲𝘀𝘁 𝗻𝗲𝘄 𝗰𝗿𝗲𝗮𝘁𝗶𝘃𝗲𝘀 𝗾𝘂𝗶𝗰𝗸𝗹𝘆. Prompt = am I testing at least 10 ads a week? → I check customer surveys. → Make sure we're targeting the right benefits and pain points (from surveys). → How good are the hooks? In 40% of the case, the hook is the issue. → Re-launch ads that worked 2-3 months ago, making variations for them. → Test static image ads (highlights, testimonials, us vs them, vector). When the performance dips, test new ads and re-launch what worked in the past. 2. 𝗥𝗲𝘃𝗶𝗲𝘄 𝘁𝗵𝗲 𝗹𝗮𝗻𝗱𝗶𝗻𝗴 𝗽𝗮𝗴𝗲. Prompt = is the page's CVR% above 2%? → Review Hotjar heatmap for LPs. → Check if the page loading speed is <2 seconds. → LP copywriting: Ask myself:    ↳ "If I visit for the first time, can I understand in 5 seconds what this product is about?"    ↳ "What's my biggest objection to buying it?"    ↳ Now, craft the message around the two questions above. If you are still driving traffic to a product page, prioritize building a custom landing page. 3. 𝗖𝗵𝗲𝗰𝗸 𝘂𝗽𝘀𝗲𝗹𝗹𝘀 𝗮𝗻𝗱 𝗯𝘂𝗻𝗱𝗹𝗲𝘀. Prompt = do I have bundles? Do my upsells bring in 10% of revenue? → Create fitting bundles if we don't have them yet. → Add the "Shipping guarantee app" for $2.99 per order. → Add checkout cross-sells. If your upsell revenue is lower than 5%, edit the upsell sequences. 4. 𝗨𝘀𝗲 𝗲𝗺𝗮𝗶𝗹 𝗺𝗮𝗿𝗸𝗲𝘁𝗶𝗻𝗴 𝘁𝗼 𝗱𝗿𝗶𝘃𝗲 𝗟𝗧𝗩 𝗳𝗼𝗿𝘄𝗮𝗿𝗱. Prompt = do we send 3 email campaigns per week? → It doesn't directly impact FB ROAS, but it does impact store-level MER. → We need to ensure we have profitable MER during low times on FB ads. → 25-30% of total revenue should come from emails. → If our open rate is around 40%, we can afford 1-2 extra emails per week. Paid ads and email marketing work hand in hand. - FOCUS on these 4 things first, as you should already have them in place. 5. (BONUS) 𝗦𝘁𝗮𝗿𝘁 𝗮𝗱𝗱𝗶𝗻𝗴 𝗻𝗲𝘄 𝘁𝗲𝘀𝘁𝘀 𝘁𝗼 𝘁𝗵𝗲 𝗺𝗶𝘅. These are the "Power 3" tests that can 2x your brand overnight: → Run whitelisting ads. → Launch an advertorial funnel. → Build an irresistible offer. Think about Domino's and their entire "*30 minutes or it's free*" angle. Or Lumin's Skincare - "*FREE product+subscription*" offer. Today, in 2024, the offer, whitelisting, and advertorials can turn your performance around 180 degrees. When performance is down, most people blame the algorithm. Try to look deeper and become one of the 1% who find solutions and reverse the performance trend. - TL;DR 1. Test new creatives quickly. 2. Review the landing page. 3. Check upsells and bundles. 4. Use email marketing to drive LTV forward. 5. Start adding new tests to the mix. P.S. What is your go-to performance solution tip?

Creative testing is your secret weapon on Meta. Here's my approach. (1) Crank Out Quality Ads - Produce as many ads as possible while maintaining high quality. Avoid creating low-quality ads just to increase volume. Have a hypothesis you want to test - whether it's a visual hook, trending audio, or a new creator - to gather learnings. (2) Track Every Ad - Use a standard naming convention to keep track of all your ads. This helps in analyzing performance over time. You don't want to look back in the last 90 days to try to see which hooks, scripts, creators, concepts, etc are your best building blocks. (3) Creative Diversity - Test various types of ads – static, video, UGC, 3D motion, AI-generated, and even "ugly" ads. The goal is to have a diverse set of creatives. They look different but also speak to different levels of awareness. Different offers, angles, headlines, scripts, etc. (4) Allocate Budget for Testing - Dedicate 10-25% of your marketing budget to creative testing. Increase this percentage if you're just starting out. Know that this will inflate your CPA but the winning creatives that come out will take your ad account to new levels. (5) Isolate Variables - Conduct 'ad set tests' to isolate variables. Ensure each test gets enough data, around 30 conversions in a week, to be more confident. If possible, you can test in lower cost GEOs, on Android, or Tiktok. (6) Monitor Spend - The ad that gets the most spend is a strong signal of its effectiveness. Keep an eye on where your budget is going. It's a good sign that this ad can scale if you port it over into your evergreen/BAU campaigns. (7) Always-On Testing - Testing should be a continuous process. Don’t stop after finding one winning ad. Keep the momentum going to discover new top performers. You need to build the muscle with constant reps. Creativity can be learned, I promise. It's fun too. (8) Volume Negates Luck - The more you test, the higher your chances of finding successful ads. This is just the reality now advertising on Meta and Tiktok. There is an exponential rise of video content on our screens. Trust me, it's going to get more difficult with AI-generated video. You need a reliable system to stand out and gain mind share. Solving this is the #1 opportunity for brands to grow efficiently heading into 2025. ___________________ P.S. - You can have both quality and quantity with Sovran

These are the battle scars of getting a business from zero to one. Raylen Margono and I launched Ando with the goal of being profitable on Meta after just $4k in ad spend (we're 100% bootstrapped). We had 50+ ads ready to test at launch so this task was extremely challenging. I had to make pause/scale decisions with far less data than I'd normally ask for. Each campaign below represents a new round of consolidation and optimization, from oldest to newest in descending order. The main levers I pulled besides scaling the most promising ads were: a) Identifying the highest CTR and highest Thumb-Stop Rate ads from the previous round that did not perform or scale and making slight iterations on text overlays, hooks, etc b) Taking learnings from our top performing ad copy (text/headlines) and incorporating that into ad hooks c) Whitelisting the most promising UGC ads directly from creator handles d) Reducing the number of adsets in each subsequent campaign, getting us down to just 1, to exit the learning phase as quickly as possible e) Holding off on launching our first Advantage+ Shopping campaign for as long as possible until it had the data necessary to perform in our brand new ad account f) using post-purchase survey data about purchase motivations/what brand they switched from to inform ad creative g) running A/B tests on LP copy in the background of all of this If our current campaign can hold at ROAS > 2, it's scale time 🚀