The maintenance of bids in Google AdWords is one of the most time-consuming activities when managing an AdWords campaign. Bid automation therefore seems to be the obvious answer, you would think. An automated system doesn’t get tired, diligently goes over all of the keywords and advertisements, is disciplined and performs its tasks in a regular fashion. Moreover, it can swiftly perform calculations that are hard to do for us humans.

Bid management isn’t easy

Since the emergence of PPC campaigns, attempts have been made to automate bids. The results in practice were often disappointing, though. Bid managers actually seemed to be steering campaigns in the wrong direction.

Often only recent data were considered, for example, without looking at predictive data. Imagine that you’re selling air conditioners. When it’s a cold summer you won’t sell that much, but as soon as the temperature rises your sales and conversion ratios will shoot through the roof. A bid manager who looks at data from the recent past may initially put in lower bids. A predictive algorithm foresees the change in weather and anticipates with higher bids.

That’s why human intuition often beat the bid manager. Tests with bid managers therefore often produced unsatisfactory results and people reverted back to manual bidding.

Importance of data

A good bid management system takes many factors into consideration. The more data you have at your disposal, the better calculations can be made. The availability of quality data is literally worth money.

Google itself also uses bid management functions. Google is able to develop very good algorithms. However, Google has two disadvantages. Disadvantage number one is that appearances are against Google; who would put his bids in the hands of a party that makes money from them? Bid management at Google is still a black box.

In addition, Google doesn’t always possess specific customer data or have knowledge of the client’s business model. Margin data is an important factor, for example. Targeting a general objective such as an average ROAS doesn’t necessarily mean that you’re also targeting margin maximization. If a product is scarce, you can also take into account whether or not you can also sell that same product through another channel. An ROAS in Google can be favorable in itself, but the ROAS in another channel may be even more appealing.

Knowledge of the business model

At ADchieve, the development of bidding algorithms is therefore a customized process. As ADchieve we contribute our statistical knowledge and automation solutions. However, first and foremost we start by listening and by identifying the business model of our clients. Based on that information, we design a first algorithm. Often, a first algorithm is basically a method that the customer already applies manually. The first step is then to automate the manual work. Gradually the algorithm is further refined and in collaboration with the clients, we unlock more and more data that we include in the algorithm. An algorithm is therefore truly a customized product and also a way for clients to structurally set themselves apart from their competitors.