In advertising, ROI is everything. If the marketer doesn’t get some sort of positive ROI on his advertising expenditures, he won’t (or shouldn’t) run said advertising. A whole side discussion can be had around how you define “return” to a marketer, which can range from very measurable “direct sales generated” to softer stuff such as purchase intent, brand lift, etc…, but we’ll leave that alone for now. And of course the marketer’s goal is always to figure how to measure and value the advertising he does, and ultimately to increase the ROI of his advertising. Take a look at the formula for how to calculate an ROI percentage:
It must not be forgotten that, according to the above formula, there are two ways to increase ROI: (1) by increasing the net program benefits, and/or (2) by decreasing the program costs. By achieving either one of these, you by definition increase the ROI percentage.
Now let’s apply this to online advertising technology. In ad tech, a vast majority of the companies I see being founded appear focused almost exclusively on the numerator of the formula (i.e., increasing net program benefits). In other words, they try to make things “smarter” by being more intelligent about which media to buy. In online advertising, that could mean focusing on bettering click-through-rate (CTR) or click-to-conversion rate, etc… It’s often easy to spot these companies, as they constantly tout algorithms as the core part of their technology and talk only about optimization techniques and “machine learning engines” that can identify the best media to buy, the right audience, etc… This is very important stuff no doubt, but it forgets that there is a denominator in the ROI formula (the cost of an ad program). In other words, there is a whole other option to increase ROI.
According to the formula, you can achieve the same ROI uplift for an advertiser as increasing a program’s benefits by 50% if you were just to decrease the inherent cost of the advertisements purchased (assuming similar/same media) by 50%. Often times, that can be a more straight-forward and fundamental problem to solve. For instance, in banner advertising that meant the creation of ad exchanges (which led to the reduction of middlemen), the reduction of waste by increased ad serving times, etc… All of these things lowered the cost of buying the same media as before, and as a result positively impacted the ROI for advertisers.
In my opinion, it often times can be a much more valuable endeavor for starting a client-performance-driven company (aren’t they all?) that focuses on the denominator as opposed to the numerator. In online advertising at least, if you build something that lowers the cost of media, you may have built something that is more fundamentally valuable to the industry and stickier than any one algorithm every could be. As one of our early advisors told us at Invite, sometimes it’s more valuable to build better piping than to worry about what water is flowing through them.
Update to post:
I’ve received a few emails and notes related to the above post, and feel clarifying one thing is important. In my post, I talk about lowering the cost advertisers pay for media as a means of increasing ROI. This is entirely accurate. However, I should have been more specific as to what I mean. It’s not by artificially lowering the revenue the publisher ultimately receives, it’s by reducing the margin and friction that occurs in the exchange of media. In the post I mentioned a few examples, such as the advent of ad exchanges, faster ad serving, the building of better piping, etc… Those examples in my head relate to making the buying/selling of media more efficient, which lets the advertiser pay less for their media but also lets the publisher get the same amount or more in revenue. I should have made that more clear.