What a week at Disney Swan Resort with Sitecore! Not only did we attend Sitecore Symposium 2018 and learn more about what is possible as well as how we can help firms get the most out of Sitecore; we also became a Sitecore affiliate partner. This affiliation will provide additional resources and support to create the optimal Sitecore solution based on your objectives and needs.
Our Partner Conference took place at the symposium in Orlando and we didn’t want to leave! There was a fantastic exchange of ideas with experts, Sitecore MVPs, and firms presenting on their successes. We wish we could have met more people and attended all of the sessions – maybe next year. With limited time, we focused on two exciting areas in Sitecore; Headless CMS and Machine Learning.
Headless Content Management
Below are a few notes we took at Symposium on Headless Content Management :
Current State of Monolithic Architecture
- Software continues to grow with features which can lead to a monolithic architecture
- Content management software is built from scratch a lot getting bigger or bloated over time
- Simple is overrated because we want more features
- A blue ocean market is fine until it becomes a red ocean to which you start to adding features to catch competitors
- 80% of users only use 20% of the features
- No one sales software anymore they just want to rent it to you
- There are always people on the edges of your software that use it in ways you don’t know. Headless can help your team to expand your edge case to fit all customer scenarios
- Adjacent market capture: You can capture another market in your software by adding features from others through headless micro-services
- Sell headless to IT as it’s more a developer sale
- Value Prop for CMO: We are going to make your IT better
Headless Key Slides
Personalization and A/B Testing are two amazing ways to customize the user experience. However, they need a solid data set massaged by data science to truly be effective. Machine Learning is a new way that Sitecore is using to provide an even more unique experience. Using user behavior to predict the right content at the right time can make all the difference between customer retention or customer loss.
Below are a few notes we took at Symposium on Machine Learning :
Recommended Providers to try or build a proof of concept
- Microsoft Azure Studio
- IBM Watson
Common Steps when starting Machine Learning
- Choose your inputs on good distributions of data. These inputs or data feeds will become your features.
- Choose a model that best fits the data science outcome you are trying to achieve. Microsoft and IBM have many predefined models to choose from and you can create your own.
- Once your data sample is chosen and you have your model, create an R language algorithm to be executed on your data
- Review and filter your data before using it within your application
- If the data doesn’t make sense make a new algorithm or choose another input data feed
- Pull data into Sitecore or your CMS using the API Tools
- If you are on Sitecore 9, you can use xConnect to do this
- Use the data you pull in on Sitecore Features
- A/B Testing
- Cortex Operations
- Real-Time Analytics your stakeholders care about
- Constantly tune your model and feeding it fresh data
- Machine Learning is not a one and done operation. Your team must constantly review, test, and tweak your algorithms to obtain peak results.
Machine Learning Key Slides