The popular belief is that performance marketing is a very well-known concept. When I look at my Clients and Traffic Providers, I can see that everyone understands this notion slightly differently (and, in some cases, mis-defines it). This results from very different methods of “practising” performance marketing and the need to approach each campaign very individually.
Let’s make it clear: performance marketing is about activities aimed at generating specific actions (e.g. lead, sales, getting users to subscribe to a newsletter). In the performance model, the client pays ONLY for these actions, so from the business point of view it would seem this is an ideal model that involves absolutely no risk and will stimulate the business only positively. But is this really the case? Can performance marketing in fact only bring benefits regardless of the actions taken, type of campaign, or the product/service category?
Performance marketing is about extensive activities divided into several stages. Depending on what kind of actions are to be generated, we define target groups. Then, we prepare the strategy and communication. After that, we search through relevant traffic sources in order to produce desired interactions. This means that we should expect a huge amount of traffic at the very first stages of performance activities. The more we search, the greater the chance for success.
At following stages, we focus on the portions of traffic that generate the desired actions. Even if the user did not complete the registration, but visited our landing page, we already have something to work with.
Through this selection and elimination of ineffective sources, we choose the target traffic to focus on. Through a variety of A/B tests (creatives, landing pages, formats, main claims, etc.), we choose the best combination of all elements that are important for the conversion.
The optimisation combines the operation of an algorithm with manual modifications of the person who is responsible for delivering specific campaign results. This two-level analysis allows us to achieve the best results and get defined actions at the assumed cost.
Are there any campaigns, then, for which performance will not be the most appropriate solution? These are some cases that are not necessarily suitable for pure performance activities:
1) The campaign is too niche. If a product or service focuses on a very narrow target group, we will probably not find it by mass searching through the whole available inventory. At best, these searches will bring “some” results, but the time necessary to reach our group may be too long and too expensive for traffic providers. The lack of interaction in a niche campaign gives it a low priority and takes it much longer to reach its target audience. Furthermore, a campaign without positive prediction usually gets lower quality traffic. Niche campaigns benefit more from customised solutions combining various activities aimed at our narrow target group. In many cases, direct activities are much more effective. For example, if we’re looking for restaurant owners in Warsaw we will probably get their information much easier by contacting them directly to explain our idea and present its main benefits, rather than by sending an email with the landing page inviting them to join our application.
2) Your target group will never make a purchase by clicking on the ad. Let’s face it: not everything can be sold on the Internet. There are goods and services (especially those more expensive, premium ones) where the decision-making process takes quite a long time. A marketer who promises to sell a luxury apartment by having a beautiful and responsive landing page is certainly a huge optimist, but also… just a dreamer.
Performance can support such campaigns, but mostly in order to obtain contact details of prospective customers who should then be handled by competent sales agents. And this should be done in person.
3) Your target group does not regularly visit Internet websites. Over 25 million Poles actively use the Internet (worldwide: 50% of the population). On average, we spend about 6 hours online! These are impressive numbers. However, this is not the whole population. If you are looking for 65+ women to whom you would like to offer a foreign language course, you are more likely to reach them through other channels, rather than by displaying ads on popular Internet websites. On top of that, if you are looking specifically for sales, you should also know the limitations of certain target groups when it comes to digital payments.
4) Your infrastructure is not ready for performance! This may sound surprising, but many companies are only now starting to conduct their business online (a good example here may be global trade, which has a hard time responding quickly to changes and adjusting to selling online). In order to be able to support your sales with performance, you must first plan all potential acquisition volumes, estimate the costs of launching the new channel, foresee back transactions (for sales), and prepare to handle additional users. Another important element of online activities is how much traffic can a given platform handle (according to the principle: performance searches huge amounts of online resources, thus providing great volumes at the first stage of activities), as well as anti-fraud protection. In this respect, it is best to use an experienced agency that has the right technology.
Performance solutions may still be used in the cases presented above but these campaigns will certainly not fully utilise the capabilities of performance algorithms. A good idea, in this case, may be a test hybrid campaign or pure performance supported by more “standard” activities, e.g. PR, word-of-mouth marketing, AdWords and the like. Testing performance solutions is practically cost-free, which is a great benefit. Therefore, I totally recommend using test campaigns for projects that have good ideas, concrete concepts and the right technical capabilities.