– ROAS (Return on Advertising Spend) is not a reliable metric for all types of businesses, especially those with repeat customers or longer purchase cycles.
– For businesses with repeat customers, the focus should be on acquiring customers at a low cost and letting their experience with the product determine their overall value.
– Instead of analyzing post-purchase behavior to determine customer value, it may be more effective to focus on converting more people and getting more products into their hands.Focus on converting more people and getting more product into their hands to increase long-term customer value and revenue.ROAS doesn’t make a ton of sense when measured as a snapshot. Here’s why: There are two types of businesses: ones where people will come back and purchase multiple times, with low barriers to make their first purchase and low barriers to come back and make subsequent purchases, and ones where people will make a considered purchase and be one and done, not likely to have a need to return for any other products. In the first situation, the goal of the brand should be to get the product in hands at the lowest possible CAC. Then let the experience with the product dictate the overall value of the customer. This whole exchange will take place over a 30-60-90 day period usually, so a snapshot ROAS doesn’t really tell you which campaign will be most successful because the bulk of the future transactions happen outside of the reportable window. In the second situation, the process for making a purchase often takes longer from an ad perspective. Research shared from larger companies stated that, on average, it takes someone 45 days on a larger considered purchase to take an action. So measuring ROAS on a limited scale doesn’t make sense, and the quality of the audience is hard to measure. The problem is revenue in most cases is first-time revenue from a customer. Now I know there are people that swear by analyzing the amount of products purchased, the types of products purchased, and everything else that happens post-first purchase to determine who’s likely to be a good customer long term (cohort analysis). I offer up this contradictory opinion: rather than chase the small percentage that has converted and build models around that, why not just convert more people? If you want more LTV, just get more product into hands. We’re running this test now where we added 50% more orders profitably in a month without massively increasing the initial revenue, but the knock-on effect of more sales equates to more repeat purchases. So this is leading to a 20% improvement in revenue mapped for this month so far, with decreased spend. Order volume is still growing. None of these follow-on sales will be attributed to the campaigns in a weekly ROAS snapshot though. So now for the fun part, what if we could create a confidence interval based on the data from a form someone filled out, the data patterns, then combine this with the products they purchased, the time between purchases, and incorporate the source of the traffic too? What if we could model someone’s likelihood of being a good customer before they purchase? Is this even possible? Not sure.Do we have the data necessary to find out? We do. We’ll be testing this soon. Stay tuned. #ecommerce #data #strategyhttps://www.linkedin.com/in/jivanco