Advanced Funnel Optimization (AFO) is an ML-based program that helps eCommerce and BFSI businesses maximize their conversion funnels. High propensity models are identified from your analytics and CRM data to create an ML model. The Winner conversion path is set for the targeted audiences to transform them into high-paying customers. Once deployed, it keeps learning and optimizing forever. As a result, your marketing strategies become smarter eliminating guesswork and rework.
Customers’ behavior on the website/app is the key to understanding useful actions. For example, purchase propensity, inquiry, subscription, and such. The AFO model will identify and attract ideal visitors to the website and personalize their experience to engage. So they can act. This model is built on your first-party data and considers your business’s seasonality, making it more pertinent to your business outcomes.
Can we get a clear understanding of user Interests & behavior to suggest products/solutions that will make difference in their life? Customer Value Maximization will help you find and identify buyer segments with higher value and intent. This will allow you to reach them on the right media, channels, and time.
Learn more about your customers, and what’s relevant to making that purchase. The third part of AFO takes into account deeper data points and provides the content or product recommendations to your customers. It is done on the basis of their content consumption or buying behavior. This is based on the following hypothesis:
- Users who interact with a category or subcategory in a similar manner share more interest.
- Users who are interacting, in the same manner, are likely to respond in the same way to the same category.
“As one of the core pillars of the annual Google JBP, the value of the customer value maximization and lead scoring propensity model has directly affected business outcomes. The ability to leverage C4M & ML for greater synergies between call center and media has led to much better overall sales conversion rates and made our media spends more ROI efficient”
Manish Dubey, CMO,
ICICI Prudential Life Insurance
“PredictN by Tatvic gave us an audience with a higher probability to submit a lead in the next 7 days. We directly exported this custom GA audience to our AdWords account and remarketed to them. As a result of remarketing to such a focused audience, we were able to achieve a 30% higher Conversion Rate.”
AVP – Head Marketing Eureka Forbes Ltd
We combine the Online behavioral data and offline data and prepare it according to the business use case and train the model accordingly. This model is built on your first party data and considers the seasonality of your business, which makes it more pertinent to your business outcomes.
Yes, the journeys for app and web are different, events triggered are different, the user behaviour would be different. thus the models will be seperate for both of them.
There is no specified algorithm, we try multiple boosting, bagging models and select the best according to the model performance.
Yes you will have the IP of the solution
High value leads, Reduced Cost of Acquisition, Optimize marketting campaign, Revenue drivers (what features impacts the highest conversions)
Yes we do have an option of doing PoC however, in PoC the model is deployed on Tatvic’s server
we can do it on T-1 basis, or hourly basis depending on the business requirement.
We even have solution for supporting real-time scoring which can be used by your SMS, Email and Call-Center campaigns. However, for real-time we will need to setup Tatvic’s Pipestream solution.