You’re Wasting Half of Your Ad Money? Yes, you. (Part 1 of 2)
“Half the money I spend on advertising is wasted; the trouble is I don’t know which half” — John Wanamaker (1838–1922)
One of the first things you learn when studying trends in data is to be weary of the following fallacy: correlation does not imply causation. In other words, just because B happens frequently after A doesn’t mean that A caused B. An observed change in the quantity or price of an item (the effect), doesn’t necessarily mean that it was caused by a similarly timed observed change in another variable (the cause). Sometimes, there can be one or more unaccounted for variables causing both observed effects.
For example, if I measured the rate of caffeine consumption by my development team as we approach a major feature launch, I would see a noticeable spike. Simultaneously, I might also likely see a drop in time spent exercising or engaging in other leisure activities during that same time period. Falling victim to the fallacy would be to conclude that caffeine caused our developers to stop exercising. Or, even worse, I may conclude that decreasing exercise causes an increase in caffeine intake. However, in this scenario we know these two events are merely correlated and have a common cause: the product deadline.
Thankfully, we just completed our launch cycle, much to the chagrin of Peet’s, Paramo, and Blue Bottle in SoMa.
I bring up this concept to showcase a related issue I’d like to call ambiguity in causation, which comes up constantly in traditional and modern advertising. Most often in investing in ads, analysts and marketers choose to invest in media to drum up interest in a service or product. The outcome is usually measured semi-scientifically using before vs. after observations. Simple cause and effect may be sufficient for some advertisers, especially those unaware of advanced analytics.
So, whether you are selling cars, mattresses, or tickets to an event, it would be hard to argue that spending more on ads could ever have a decrease in the desired output such as sales or product awareness. However, measuring the return on investment, even at a coarse level, requires us to look at data causation in more detail (aka work). This is why the cynical or lazy approach is to argue that ads are a waste of money, or they have a neutral or negligible effect on outcome. That is unfortunate but incomplete truth. Not all campaigns will be successful, and even with the knockout successes, results and data are usually shrouded in ambiguity.
Let us take a deeper look at the latter case. Say you book a huge advertising win for your company and conclude that your endeavor clearly won you some increase in new business, but you may not know which of your specific efforts did the trick. This uncertainty poses a challenge when you want to repeat your successes in the form of follow-on advertising, explain why you were able to achieve what you did, or repeat your success in a slightly modified scenario, perhaps a different region.
The Advertiser’s Dilemma
John Wanamaker’s famous quote highlights very clearly the Advertiser’s Dilemma: investing in ads, even on successful campaigns, leaves ambiguity about the effectiveness of the specifics. If I purchased 20 equally-priced billboards throughout Manhattan, and saw a huge uptick in web traffic to my site, I’d certainly be bragging around the dinner table about how my awesome purchase caused huge success. However, if I had to cut down budget to only 10 ads, which ones should I cut? Without a way to track down to the effectiveness of the specifics, I’m left guessing.
Fortunately Lawfty doesn’t buy bus ads, radio spots, or billboards like a lot of big legal advertisers. We focus purely on what is trackable and data-driven. With online ads, our dilemma is a manageable one thanks to location and time-of-day targeting, and the ease with which digital advertisers can attribute online activity to specific ad spend sources.
That’s not to say that there isn’t a lot of work that goes into acquiring/formatting the data, optimizing bid schedules and reducing the ambiguities. I am also not claiming that, despite our best efforts, ambiguities don’t remain. They certainly do, and in Part 2 of this series, I’ll describe some of these challenges in more detail. I’ll focus in on how we keep ambiguity in causation in check, which is just enough to allow us to build out a scalable system for channeling demand for legal services to the right place at the right time.
Originally published at theread.lawfty.com on April 5, 2017.
