This year some of your customers will make purchases through websites over the internet. Others will also buy from your brick-and-mortar stores. But many of your customers will buy from both. It is crucial that you are on top of the game and have the right measurement mechanism when your customers are shopping through different channels.
There are many stages involved before a customer actually buys something online. Let us say he or she sees an ad that looks interesting. Not everyone will click on the ad right away, and some people will do it only after the ad has appeared four or five times. The ad alone does not convince all customers to make a purchase. Some customers need more information and will Google to learn more. It doesn’t matter how much time it took between seeing the ad and making a purchase; many channels assist in the process of decision-making.
For successful online advertising, marketers need precise data about buying patterns that convey up-to-date and appropriate information. Such data directly impacts decisions about advertising drives and are useful as guidance tools for expansion efforts. With the use of new techniques and technology, much progress has been made in digital marketing. However, it has become even more difficult to get data on where, how, and if conversions have taken place. This fact is particularly true with cross-device conversion (where customers engage with the brand on multiple platforms and devices) and view-through conversions (where customer conversion occurs on a brand’s website after seeing an ad but not clicking on it). The reasons for this difficulty are many; there could be cookie restrictions, lack of coherent information coming from multiple devices, or even due to the use of different browsers on the same device. It is in this kind of environment that Attribution modeling is required.
Attribution modeling is the procedure that digital marketers use to ascertain the importance of different advertising strategies. It gives weighted credit to all the touchpoints that customers go through before a successful conversion or sale.
Understanding the Attribution conversion model
Let us take the example of Google Ads. It considers website conversions that should ideally be attributed to clicks and views on Google ads but often doesn’t get noticed. It is done by using statistical techniques. It first identifies traffic with conversions that are easy to track and then applies the traffic patterns with missing or incomplete data. This conversion model uses data that has been recorded like date and time, types of devices, the number of clicks on an ad, keyword searches, and others. Using all this information, a model is produced of consumer conversion patterns so that even unnoticed events that lead to conversions can be recorded.
Campaign Manager 360, an online ad management system for advertising agencies that manage digital ad campaigns across different channels uses the Attribution Modelling Tool to give credit to different web traffic streams. Since September 28, 2020, all attribution models were redirected to conversion modeling for Display & Video 360 and Campaign Manager. This change was made to improve the tracking down of missing conversions.
The tool offers a variety of attribution models that assign more or less credit to different touchpoints:
- Floodlight Model: In the floodlight model, conversions are attributed to website visits after clicking or viewing an ad.
- Last Interaction: In the last interaction model, 100% of the conversion is attributed to the last channel of interaction before the consumer made a conversion.
- First Interaction: Here, the conversion is considered entirely due to the customer’s first point of interaction.
- Linear Model With this attribution model, credit, calculated in the form of revenue generated from the deal, is split uniformly between all the interactions that happened before the deal took place.
- Time Decay: Most of the credit is given to the interactions closest to the conversion. Credit is assigned using a half-life of seven days. It means an interaction that took place within seven days before a conversion (say a social media post) gets half the credit of another interaction (an ad click, for example) 1 day before the conversion took place.
- Position-Based Model: This model is also called U-shaped attribution. 40% of the credit is given to the first interaction; another 40% is attributed to the interaction converted to lead creation. The balance of 20% credit is given to all other interactions between them.
- Social Model: Just like the linear model, attributions credit is given uniformly with two factors being considered.These are:
- Impressions(I) = number of people who have seen the content online
- Social Engagements (SE) = mentions and conversations that pertain to the brand or products or ads and others
Different weights are given as below:
- I with no SE (seen an ad on Instagram, but done nothing) = 0.5 times weightage
- I with low-value SE (e.g., Seen ads on Instagram and clicked, but not shared) = 0.75 times weightage
- I with high-value SE (e.g., Seen ads on Instagram and mailed coupon to a friend) = 1.5 times weightage
Depending on the goals of the marketing strategy and the business involved, one can choose any of the attribution models. What works best for each marketing channel can be understood by looking at it against every attribution model. Every one of the models above has its advantages and disadvantages. For example, if your website is used for selling commercial properties, it probably doesn’t have a quick conversion cycle. The “Last Interaction” attribution model for conversion may not be the right choice. On the other hand, if you only want to know from which channel your customer made a purchase (for example, buying movie tickets), the Last Interaction model would be just right.
Besides the conversion attribution models described above, Google uses the Data-Driven attribution mode. In this model, past data is analyzed using algorithms to anticipate which of your channels or campaigns or even keywords will drive higher conversions. The model labels samples of ad interactions that give the best results, hence helping marketers identify where they should concentrate their marketing efforts.
It is good to remember that any attribution model only applies to the conversion action it focuses on. When you set up your conversion action, you can choose one model and later change it to another model to decide which one works best for your business goals.
Gaining insights into how potential customers interact with your marketing efforts, whether ad content or visiting your website, can give you important information. You can then run your campaigns to make them more relevant and appealing to your target consumer. It will automatically lead to an improvement in conversion rates. While learning the theory behind a particular model may sound perfect, you won’t know the true impact before you try it out. It’s a good idea to experiment with various conversion attribution models and allow your strategies and marketing ideas to evolve. Working like this will help you build a sustainable model and enhance your brand’s value while taking care of data security and your customer’s need for privacy.
We hope the article above was detailed enough for you to gain a good perspective of conversion modeling. With all the information you have now, you can now scrutinize your current models and update the settings. If you still have some lingering doubts and need more help in helping you realize your digital marketing goals, we are here to offer a helping hand. Do get in touch with the Full-Service Programmatic experts at Hopbug Media. We will help you choose the model that is most beneficial for you.