Kärcher is a German company founded in 1935 and is currently a leader in cleaning technology – and sell everything from high-pressure cleaners to military decontamination equipment.
It also carries out big cultural sponsorship programs where it cleans up prominent buildings around the world such as the London Eye, Brandenburg Gate, Rio’s Christ statue, St Peter’s Square, and the Statue of Liberty.
In order to elevate how they communicate to consumers and provide a cleaner and more personalized interaction, Kärcher decided to upgrade their content management system. Their main limitation was the lack of a content storage that was separated from distribution, with simple access for their editors.
A second challenge was their CMS didn’t hold flexible content types that could be fit for more unconventional distribution channels. And finally, the lack of ability to update their content live in various languages.
And to drive greater customer engagement, Kärcher plugged in their content to Contentful when developing voice control for Alexa.
Content models have to meet the standards of the voice provider as well as remain easily accessible to the editors of Kärcher. Because the problem is that if models aren’t built properly, Alexa is unable to understand the prompts and can’t return the requested info.
The complexity expands when you need to publish content in more than one language as it’s not only related to creating more content but managing it and updating it.
So what Kärcher did was place its content into the content management system of Contentful. This resulted in the ability for their customers to access info and interact with the brand via voice control, and give each customer a unique experience.
Building with Contentful and Alexa
There were a series of reviews that the team went through to reach the decision to go with Contenful. It ultimately came down to its flexibility, user-friendly interface, and security features – all key aspects of a modern content management system.
The first step was to create a proof of concept and then went on to build a content model for the actual, full-scale project. Next up, Kärcher editors and developers worked together to connect its GraphQL content APIs to those of Alexa, enabling editors to handle the voice content through Contentful’s custom content model.
Then Kärcher’s editors worked through its content, rewriting and rewriting and then condensing. Because for voice assistance, the performance is better if the statements are brief – think about it, it’s not that easy for bots to listen to long statements as they’ve more complex.
Iterating was smooth once the content was stored in Contentful. User testing followed through a couple more iterations that turned beneficial. As one example, they discovered that slot values can only be used one time. So using the same word various times causes the Alexa system to stall.
Cleaning tips through Contentful to Alexa
Kärcher has created two Alexa Skills through Contentful. The first one: customers can control the irrigation for one of Kärcher’s products, the Watering System Duo Smart Kit.
Scaling is another feature that Contentful enables Kärcher for when it comes to localizing content. Scaling across languages and markets. Today, it’s available in English, German, and French. And their content can be reused for their chatbots as well as their IoT machines.
“The integration of Contentful reduced the complexity of our architecture,” said Jens Hildenbrand, program portfolio manager of Kärcher, “due to the excellent API and technical documentation in combination with the convenient backend for data modeling.”
The core takeaways that Kärcher found was that they could repeatedly test including user feedback on products to better understand what products their customers are looking for as well as how they’re using them; they can use the same content structures for their apps, websites, and chatbots, squeezing the length of time it takes to create content; and the reduction in barrier between the editors and content, so that any content changes made can be done by editors themselves.