This post is part of a series of case studies that discuss solutions implemented by Syntelli’s data science team. To read more of our success stories, visit here.
Client: Large Manufacturer of Home Appliances
One of our customers is a manufacturing company producing several types of home appliances. The manufacturing company was interested in better understanding how their marketing channels (direct, online display ads, radio and tv ads, billboards, and others) are contributing to their sales. In addition, they were interested in the effect of each website page (such as product information, Where-to-Buy, FAQ, etc.) on the sales of different lines of product.
Client Challenge: Omni-Channel Marketing Attribution
The common solution for this kind of problem is using the algorithms of attribution modeling which exists in Google Analytics engine. For this aim, marketing data analysts define a conversion metric as the number of consumers who landed on the Where-to-Buy landing page. Then, the built-in engine of attribution modeling in Google Analytics provides the effect of each marketing channel on the defined conversion metric.
However helpful, this straightforward solution was not enough for our customer. Not all the consumers who buy an appliance view the Where-to-Buy landing page. In addition, the attribution modeling of Google Analytics evaluates only the channels used before conversion, not the website page that the consumers could have viewed before their purchase.
Moreover, consumers might view some web pages after purchase, e.g. for downloading a manual or registering an appliance. Thus, a strategy had to be adopted by Syntelli’s data science team to consider these challenges and provide a realistic model between sales and consumer’s website activity.
In the first step, we extracted the website activity of consumers which was tracked by Mixpanel. Why use MixPanel instead of Google Analytics? While Google Analytics only tracks the web pages, the Mixpanel tracks the actions people take on the website. Mixpanel tracks all the activity a person does on the website such as viewing a page, registering an appliance or downloading a manual, and assigns all of them to a unique id.
Our attribution modeling needed to navigate how the customer was exposed to media and site content – and in what order the exposure occurred. In essence, we had to have our data match their online journey. Fortunately, we were able to use a graph database instead of a more traditional relational database system.
Graph Databases are a relatively recent additional to the Data Science tool kit. It allows us to not only define relationships between our data, but also assign attributes to the relationships themselves. When we add ‘time’ as one of those attributes, it makes simple the once complex task of mapping in-site user activity.
For our graph DB, we implemented it in Neoj4. All the information of Mixpanel were stored in the designed graph database. In the next step, unique IDs and purchase dates of people who bought at least an appliance were extracted, and channel and web pages that buyers had viewed before and after purchase date were listed.
Now that we had our data structured in a graph, we were able to apply a more comprehensive attribution model page views and channels before and after purchase date.
Using a method known as Bayesian Attribution, we developed a probabilistic graph of all the different media and site content journeys leading to a sale. Using this graph, we determined the reduction in probability of sale by leaving out a particular step. The change in probability gave us the value of a given media channel or piece of content in the journey – i.e. its sales attribution.
While our enhanced attribution modeling method provides useful information for our customer about the effect of marketing channels and web pages on sales, other interesting information about consumer behavior was also extracted. For example, percentage of consumers who used Where-to-Buy page to purchase, time between page visits and purchasing, as well as patterns in purchasing different appliances by a consumer. The manufacturing client used this information to improve their website structure, build their recommendation system and optimize their advertising campaigns across multiple (offline and online) channels.