Best Practices for Conversion Rate Optimization Testing

I've audited 100+ brands all with the same goal: profitable growth.  80% don't have a structured testing framework. Here's the one we use to take brands from 6 figures to 8 figures. 👇 1. Establish your monthly hypotheses. Break these down into: 🔹Creative 🔹Landing Page 🔹Ad copy I wouldn't bother much with audiences these days. 🔑 2. Run at least 8 major tests per month. Think BIG swings (net new concepts, very noticeable differences between iterations A & B) And, sprinkle in 1-2 hyper-targeted tests (small iterations of top performers) per week. 3. Based on this, attribute a % of spend you are comfortable testing with. We recommend 10-15%. This is enough to get learnings fast, but not too much to rock the boat. (Caveat: some brand owners are more aggressive and are comfortable with a higher % of spend) 4. Now you need to outline your testing parameters. 🔹Know when to scale 🔹Know when to cut 👉You should know within 24-48 hours if the tested material is going to be a winner or not. *This is a GENERAL guideline, it may take longer with high AOV or long path-to-purchase products/brands.* 5. Set your daily budget at 5X your average CPA So if your average CPA is $60, set your budget at $300. Within 24 hours you are going to get some strong early indications of whether your hypothesis is correct or not. 6. So you're 24 hours in and now it's time to assess how your test is going. 🤜if CPA <25% over your CPA target, keep running for another 24-48 hours 🤜if CPA >25% over your CPA target pause the test unless secondary metrics are performing. ✨You may keep this test running when CPA >25% over your average if: 🔸 Cost per add to cart is 5-25% above target 🔸CTR 25-50% above target 🚦When CPA is >25% your average or target, this is your flashing light to scale. Here we like to be aggressive, but moderate. We increase spend 25%-50% daily based on CPA and secondary metric performance. 7. ANALYZE your results weekly to drive quick pivots, and monthly to plan your next round of tests. Note: We use advanced ad-naming that feeds directly into Motion reports. There, we can compare things like: 👉 Messaging Angle 👉 Hook Text 👉 Hook Visual 👉 Sales Sequence 👉Consumer Psychology Pillars We can track this all across multiple creative tests. It's important that you're collating data around major creative testing themes or "buckets" each month so you can get the 10,000ft view of key trends. This will drive the future "big swings" and get you OFF the hamster wheel of tiny meaningless iterations. What does your approach to testing look like? ❓

Q: Which metrics matter most? Marketers are drowning in data these days, and not all of the data is equally useful or insightful. When we're managing performance campaigns for our clients, we analyze their business, their product and their audiences to understand which key metrics are going to help us move the needle. Our goal is not simply to know what worked but why it worked - so that we can make it work even better. A great example is multi-touch lead attribution. As marketers, we love to know how many website pages or content pieces did my prospect view on average before converting. But what is the information telling me? Is it telling me that something about the behavior of my prospects, or the effectiveness of my content? Did they dig deeper on the website because they were interested, or because they were confused? Before you set up multi-touch attribution, think about what you're hoping to learn, and why. Perhaps your goal is to reduce the number of touches required to lead to an action, so you can generate more leads with less effort - that's a worthy goal, and something attribution data can help you achieve. If you don't know where to start with your marketing data, start with a vigilant focus and curiosity around conversion on high-intent pages. "Get a demo" or "sign up" or "pricing" are critical pages on your website. Next to your home page, no page is more important. You should be testing it constantly and iterating quickly to learn which changes to those pages move the needle on those actions. Test headlines, CTAs, forms and offers. Try a form submission against a Calendly link and see which action leads to more meetings. Test a green button against an orange button, "Get Started" against "Get the Demo." A modest improvement to your conversion rate on a high-converting page can results in a huge increase in leads and revenue. Whatever you're measuring, make sure you're taking a step back to think through what you're hoping to learn, achieve, and do with the data. Analytics are a great tool, but you're the master craftsman. Swing that hammer with skill! ⚒ 🚀

A DTC SAAS company came to us with limited ROI on paid ads. We revamped their paid campaigns to get them: - 6X growth in 6 months - 37% reduction in CPA - Achieved net profitability in Series A This thread will tell you ALL the details of how we did it: One of Spotlight’s specialties is Health & Wellness. This telehealth business that had serious ambitions for growth, but struggled to figure out social media. We started with our traditional 3 hour diagnostic. Broke down target market, content strategy, net margins, KPIs. Everything. It was important to have our team’s goals match the client’s. Once incentives were aligned, client & agency teams worked as ONE INTEGRATED TEAM. Slack channels, Monday boards & sprints set for success. We started with set-up. Best-practice Meta CAPI set-up & integrated with the client’s software. Each conversion action in-platform matched back to corresponding Meta Events. The plan was to spend big, so we wanted to spend extra time on the foundations. Creative strategy came next. We wanted to focus on high-volume creative testing using a hypothesis-led approach. We had a 2 hour brainstorm as ONE TEAM. Mapping our hypotheses for why the target customer would use our telehealth service over competitors The Spotlight team then turned these hypotheses in to funnels. Designing ads, landers & post-purchase emails specific to each hypothesis. We set-up 3 hypotheses in parallel, each with its own UTM parameters It was time to run ads A week in, we’re starting to see divergence. All funnels are converting, but at different CPAs We map these back to hypotheses — and determine which educated guesses were right or not We prepare our next hypothesis-based funnels based on these learnings For example. We have three hypothesis-based funnels 1. Convenience 2. Price 3. Quality We learn that the Convenience angle gets $30 CPA, but Price has $44 CPA & Quality $65 CPA We note that Convenience matters more than Price & Quality for this target market We start testing the next 3 hypotheses 1. GenZ Angle 2. Parents/Boomers Angle 3. Elderly Angle Importantly, each of these is laden with Convenience factors RELEVANT TO EACH SEGMENT What counts as “convenient” for Gen Z isn’t the same as for Elderly We soon learn that Elderly converts at $25 CPA, Parents at $37 CPA and GenZ at $52 CPA So we ask ourselves Why Is Convenience So Important for Senior Citizens?! You see this question sparks more hypotheses For example, is it because they can’t drive? They need a carer to take them? So you see, this system is self-learning Your ads become proxies for testing your educated guesses And if you’re doing your job right Your educated guesses are getting better & better We did this process 10X a month for 6 months. And the results? 37% reduction in CPA From $60 to ~$37

How I find conversion rate opportunities by breaking down the shopping funnel: Instead of looking at your entire funnel conversion rate (2-3% on average)... Step 1. Break it into parts. 1. All traffic 2. Non-bounce (% Sessions viewing 2+ pages) 3. Product Viewers (% Sessions viewing 1+ product) 4. Add to Cart (% Sessions adding 1+ product to cart) 5. Checkout Start (% Sessions starting checkout) 6. Checkout Complete* (% Sessions completing 1+ orders) *You can also break down the checkout flow further: Billing/Shipping > Review > Thank You As a percent of the total, a typical e-commerce site might be: 1. All traffic: 10,000 sessions - 100% 2. Non-bounce: 7,000 sessions - 70% 3. Product Viewers: 3,000 sessions - 30% 4. Add to Cart: 800 sessions - 8% 5. Checkout Start: 400 sessions - 4% 6. Checkout Complete: 300 sessions - 3% Step 2. Calculate the % moving to the next step The KEY is to look at the conversion rate between steps. Calculate by dividing the sessions on each step over the sessions from the previous step. 1. All traffic: NA 2. Non-bounce: 7,000 / 10,000 = 70% 3. Product Viewers: 3,000 / 7,000 = 43% 4. Add to Cart: 800 / 3,000 = 27% 5. Checkout Start: 400 / 800 = 50% 6. Checkout Complete: 300 / 400 = 75% Step 3. Look for trends You don't need to worry about ecommerce benchmarks. Your marketing channel mix, product type, and audience will all influence your numbers. Focus on YOUR numbers. This is your baseline. Trend these rates over time, and watch for anomalies. Step 4. Improve each step methodically Does your checkout completion rate look low (75%)? Maybe consider: - Checkout Form optimization - Adding new payment types - Simpler discount codes - Accurate delivery estimates Is your Add-to-Cart rate low (27%)? Maybe consider: - Pricing optimization - Additional social proof on PDP - Improved product images and videos - Digging into inventory and availability Step 5. Track your results As you make improvements (or run experiments) measure your intra-funnel rates. It's much easier to track improvements compared to looking at your aggregate conversion rate. Are you breaking down your e-commerce funnel? #cro #conversionrate #ecommerceanalytics

THIS IS HOW YOU A/B TEST AS A PRO MARKETER 👇 Most people think A/B testing is just: “Try two ads and see what wins.” But that’s amateur stuff. Pro marketers test with purpose. Here’s how they do it: Step 1: Start with a clear hypothesis Not “Let’s see what happens.” But: ✅ “We think testimonial-style ads build more trust than product demos.” ✅ “We think shorter hooks increase CTR.” ✅ “We think static images outperform videos for this angle.” Step 2: Test ONE variable at a time Want clean data? Don’t change 5 things. Just test: → Hook A vs. Hook B → Offer A vs. Offer B → Creative format A vs. format B One change = clear result. Step 3: Know your sample size Don’t kill a test in 24 hours. Let Meta collect enough data or you’ll just be guessing. 💡 Tip: Wait until you have 50+ conversions per variation before making a call. Step 4: Look beyond ROAS Check: • CTR • CPM • CPC • CVR • AOV • MER (blended!) Sometimes, an ad with lower ROAS is building top-of-funnel momentum. Zoom out. Step 5: Scale what works fast Once you find a winner: → Make variants → Build new angles off the insight → Feed it back into the system That’s how pros compound growth. A/B testing isn’t random. It’s a system. And if you do it right You never guess what to make next. You already know.

Dynamic pricing is an effective tactic to increase conversion and revenue for subscription products. When I first tested dynamic pricing while leading subscriptions at TechCrunch, we were able to increase conversion rate by 22% while also increasing 1-year estimated LTV. Here's how we did it: 1️⃣ Identify what impacts conversion We investigated which variables were most strongly associated with conversion, and we found 10 variables (see the first image). We then used a machine-learning algorithm to score all users from 0-100 based on the criteria. 2️⃣ Create marketing segments We used the scores to create marketing segments based on the likelihood to subscribe score. We could have created 100 segments, but that’s overly complex for a first test so we simplified it into three groups to reduce scope (low, medium, and high). We referred to the score as the LTS score, or “likelihood to subscribe” score.  3️⃣ Develop hypothesis and run an experiment Our hypothesis was that segmenting with price differentiation would lead to a higher conversion rate and higher LTV than a static experience. We ran an experiment where users with a medium and high likelihood to subscribe score received a higher trial price point ($5 first month), and users with a low likelihood to subscribe score received a lower trial price point ($1 first month). See the second image for the test plan. 4️⃣ Analyze the data We looked at conversion volume, conversion rate, and gross revenue, and then modeled the estimated LTV for 1 year. Revenue and LTV numbers are intentionally removed from the image for LinkedIn sharing. Shown in image 3, the results were:  *Using dynamic pricing led to a 22% lift in conversion and higher revenue than a static paywall experience.  *Conversion rate for the medium and high score segment was 2.5x higher than the average of all other segments. The test was initially a success. It also created ideas for follow up tests and analysis. Some of the smartest subscription businesses take a similar approach. For example, The New York Times uses a machine learning algorithm to create a "dynamic meter." Every user gets a slightly different experience with the meter in order to optimize and balance engagement and revenue. Are you taking advantage of dynamic pricing to optimize revenue for your product?

One of the most painful lessons I learned early in my CRO career was this: not every win is scalable. We once ran a homepage test that produced a 28% uplift in conversion. The metrics looked great. The client was thrilled. But one week later, their operations team was drowning. Fulfillment delays, inventory shortages, and customer support flooded with “Where’s my order?” tickets. It wasn’t the test that failed. It was the system that couldn’t absorb the outcome. When we dug in, we found no stress testing had been done pre-launch. No one had asked: - What happens if this wins too well? - What part of the business gets overloaded first? That changed how we approach every engagement since. Now, before every high-impact test, we walk leadership through “if this works” scenarios, not just on UX or metrics, but on fulfillment, staffing, and downstream impact. Testing isn’t just a marketing function. It’s a pressure test for your entire operating system. Because growth doesn’t just come from lifting numbers. It comes from preparing the business to carry the weight of that lift. #conversionrate #conversionoptimization #cro

In the past 5 YEARS, I've delved deep into the world of A/B testing in e-commerce, and here's what I've discovered: The reality of today's digital landscape is that A/B testing is not a luxury; it's a necessity. Here's your action plan: ☑ Step 1: Embrace the Art of A/B Testing    - Recognize A/B testing as the process of comparing two versions of a webpage or funnel    - Understand it's like choosing between two stores: one that's vibrant and inviting versus one that's dim and disorganized    - Realize that in e-commerce, every click, second, and pixel counts, and A/B testing is the key to optimizing these elements. ↳ The importance of informed decisions. ↳ The power of customer preference. ↳ The necessity of continuous refinement. (but there's more to it) ☑ Step 2: Implement A/B Testing on Your Funnel    - Set up two different versions of a single funnel element, like a landing page or an email campaign    - Test elements like design, content, and calls to action to see which performs better ↳ Crafting compelling digital experiences. ↳ Learning from real data. ↳ Adapting strategies for maximum impact. (Use these insights to refine your digital strategy) ☑ Step 3: Analyze and Interpret the Results    - Track and compare the performance of each version    - Look at metrics like conversion rates, click-through rates, and customer engagement ☑ Step 4: Optimize Based on Findings    - Use the insights gained from A/B testing to refine your digital strategy    - Continually update and improve your funnels based on data-driven decisions ☑ Step 5: Experiment with Different Elements    - Test various aspects of your funnel, including headlines, images, and layout    - Don’t be afraid to try bold changes; sometimes, the smallest tweaks make the biggest difference ☑ Step 6: Continuous Learning and Adaptation    - The digital world is ever-evolving, and so should your strategies    - Stay updated on trends and continually incorporate new insights into your A/B testing ☑ Step 7: Share Your Learnings and Successes    - Document your A/B testing journey and share your findings with your team or audience    - Use your successes and learnings to position yourself as a thought leader in digital marketing And that's how you refine your digital strategy through A/B testing! It’s not just about testing; it’s about strategically enhancing every aspect of your digital presence. P.S. This guide is for those who want to transform their digital strategy through data-driven decisions. But if you’re comfortable with the status quo, You might stick to the routine. The choice is yours. Innovation or Tradition? What’s your take? Drop a comment and let’s discuss :) #ecommerce #abtesting #revenue #roadmap

Two simple but impactful wins for your website. ➡️ Show, or better yet, interact with the product and let your prospect book a call immediately when they are ready. Two common things I often recommend to increase your website pipeline that often get INSTANT objections. The objections are all internal issues. My response, "That's an internal issue. Let's align with the right people and start moving the number." Sometimes you just need to remind people who owns the number and we're all matching to move it together! Next, I use data to show them the potential impact. I pull previous results we've seen but I also reference external benchmarks. 1. Calendar booking to increase conversion rate. Chili Piper just launched a great benchmark report. "86.97% of form submissions are from qualified customers. And 65.09% of those qualified prospects are booking time with sales. Pretty good compared to the baseline of 30-40%" I model this out using their data and show them how this increase can impact revenue. 2. Let people see and interact with your product BEFORE they talk to sales. The best way for your buyer experience is to do an interactive tour. "Interactive tours have 2.5x CTR when compared to videos" Source: Navattic State of the Interactive Product Demo 2024 Video is still a win in my book but it's a half step. If you are serious, take the full step. These benchmarks help show what can be possible based on what others are doing. Model it out for yourself and input your adjustments based on your business but TAKE ACTION. P.S. Not a sponsored post. #gtmstrategy #websiteoptimization #productdemo

A lot of brands focus all their energy on the ad—but drop the ball completely in the post-click experience. This is a great way to light your ad spend on fire. To maximize performance and cohesion in the customer journey, you’ll want to use test these three 3 strategies: 1) Build 5-Reasons Why Listicles 2) Create Comparison Articles 3) Correlate Your Upsells To expand on each… 1) Build 5-Reasons Why Listicle You should absolutely test a 5-Reasons Why Listicle after your customer clicks on an ad. This is where the purchase decision gets made. You should serve key information necessary to convince your customers to buy immediately upon landing. Think of it as a PDP masked as an article: You need to enable the person to purchase inside of that article because they are not reading the ‘5 reasons why article’ and then saying, “Oh, I really need to read the PDP now.” 2) Test Comparison Articles This is where you service shoppers a comparison between your product and a popular competitor. When highlighting the key differences, remember that you are creating the shopping moment there and then. Be as informative as possible, covering what might’ve been most vital in a PDP. Don’t make customers read both, or they’ll fall out of the funnel. 3) Correlate your upsells Your upsells should correlate to the primary product flow that originally engaged a customer. Don’t drive customers away by pushing products that don’t meet their preferences. Finally, the hierarchy of upsell products you are promoting matters just as much. Test this order regularly to see which variables create a material change in conversions. I continue to see these crush for brands using FERMAT, and just in general looking across the ecom space. If you want advice on multivariable experimentation, head over to my page and check out my Whiteboard Wednesday video from March 14th! PS: if you have any questions, feel free to drop them in the comments!