Why Data Validation is Critical for Bid Management
It’s no secret that marketing is evolving. New media, new models and a changing focus have signaled a shift from broad campaign measurement to per ad ROI. Better tracking at a click/impression level has enabled this focus on maintaining margins, planning budgets and generating sales. Now more than ever marketers are extremely focused on using statistics and all that data to drive their decisions. Tag management platforms not only offer a great way to keep up with this changing environment which enable marketers to quickly test new media and new optimization vendors, but more importantly they can be readily deployed to monitor the ever critical data streams those vendors require.
As digital marketing channels have become predictable, many organizations now depend on them for bottom-line success. Most marketing teams have sales numbers directly tied to the company’s revenue forecasts. The days of casting out brand marketing with broad messaging and hoping it generates visitors is over – it’s now about $1 out and $2 in. This performance marketing has been further fueled with the growing availability of auction-based media, such as PPC, display, re-marketing and social advertising. These formats are so popular because of the tremendous volume they provide and the advertiser’s opportunity to beat the system and generate amazing returns. It’s because of this fundamental shift in the marketing ideal and these new competitive landscapes that bid management solutions are now a fundamental component of the digital marketing strategy.
What’s a Bid Management Solution?
For those of you not familiar with bid management solutions, they routinely monitor the performance of digital advertising against its cost and automatically adjust media spend via bids to ensure that all media spend ROI is positive. These systems are absolutely reliant on data, both from your web properties, representing returns/revenue, as well as the corresponding publisher, representing the cost of media. A hiccup in either of these data streams can be disastrous. These solutions use algorithms that are focused on protecting advertiser investment, which means that if they begin to report costs without returns, the first action is to reduce costs by way of dropping bids. Ultimately if there is a problem, this can cause huge missed opportunities to capture traffic and drive sales. Luckily, the best bid management solutions leverage learning algorithms that can account for fluctuations by weighting historical performance when making day-to-day decisions. This is a great feature because it can handle seasonal fluctuations and the occasional spike in data.
Where It Gets Tricky
There is, however, a dark side to this story. Most bid management solutions operate one day behind, meaning that every morning they aggregate data from the previous day from the publishers via APIs and the data that they have tracked via their own tags. For example, PPC bid management solutions will typically take the aggregated cost of a keyword downloaded from Google and compare it with the sales created from the traffic generated from that keyword. The net is the ROI and that is then used in the algorithm for the following day. In simple terms, if the ROI is positive, the solution will keep the bid or potentially increase it. However, if the ROI is negative it will likely decrease the bid. The tricky part is that the bid management solution doesn’t know what it doesn’t know. Let’s say in a recent release that someone forgot to map the new order total value to the bid management solution tag (code snippet in the web page). The next day, all that traffic will just appear to be negative ROI. Considering the reports are a day behind, it can take at least day to figure out when something has gone terribly wrong. From there, it can be days until you know exactly what needs to be fixed. In the meantime, the damage has already been done. Keep in mind that many of these algorithms rely on historical data, so fixing the tag is just the beginning. Depending on the publisher and the bid management solution, it can take days or even weeks to recuperate from the effects. Since returns are tied to spend, the algorithms are slow to increase bids and ad spend based on the perceived poor historical period.
Validating at the Point of Collection
The challenge above is dramatically impacted by the time it takes to identify and remedy the issue. Instead of a data spike, you end up with a curve and that curve looks like a normal trend vs. an outlier to the algorithm, and is further taken into all future decisions. Some bid management platforms offer alerts and tools to help avoid these scenarios, however, tag management solutions (TMSs) offer the most comprehensive protection. TMSs are at the point of collection, and can employ data monitoring that will check not only that a tag is properly executing, but will also make sure that the data is being properly set. In the scenario above, the tag was loading, however, no order total was being properly set. A TMS can set a condition that checks for the order total to be greater than $0 anytime that the bid management tag loads. If this condition is not met, the TMS will email the advertiser immediately. The advertiser not only knows there is a problem before the reports show the issue, but the problem is also identified in detail. This combined with the ability for the TMS to map the new order total value from the page directly into the bid management tag without involving IT is an unparalleled advantage. The user is able to identify and solve a major problem within minutes that would have taken at minimum a day to even discover. This data will then look like a spike instead of a curve and the algorithms will not miss a beat.
They say that you are only as strong as your weakest link, and in marketing that is data. The right tag management solution will give you data validation that is essential to reinforce that link.