Because the COVID period started and prevented folks for a protracted time period from eating in at eating places, shoppers in all places have more and more relied on restaurant ordering and supply apps to place meals on the desk for themselves and their households.
To handle the shake-up in food-consumption dynamics, Yum! Manufacturers’ digital and expertise groups invested considerably within the growth or enhancement of such apps for our eating places, together with KFC, Pizza Hut, Taco Bell, and The Behavior Burger Grill.
For KFC-United States particularly, the idea of getting a restaurant ordering app was comparatively new. To encourage KFC clients to obtain and use the app, we wanted to make sure that it was “related, straightforward, and distinctive”—or, RED, as our earlier CEO, Greg Creed, appreciated to say.
However to actually make sure that it was RED, we wanted metrics. We would have liked to know if the app was certainly making the method of ordering fried rooster simpler. Have been folks happy with the app? Have been there recurring patterns amongst clients who liked the app (or didn’t love the app)? Did sure app launch variations carry out higher than others?
These have been among the many questions we needed to discover solutions to. Though each Apple and Android present entry to client rankings and opinions, they don’t present a deep dive into what opinions imply for a product. So, we turned to Domo, and the software that has turn into our secret sauce: Jupyter Workspaces.
Jupyter Workspaces provides us the flexibility to entry and analyze this qualitative information. In my expertise with different enterprise intelligence platforms, textual content evaluation has been restricted to phrase counts and phrase clouds.
Pattern of a Domo/Jupyter Pocket book mission carried out on Doordash Opinions
Jupyter Workspaces, alternatively, takes textual content evaluation to the subsequent degree, permitting practitioners to mix Python’s superior Pure Language Processing (NLP) capabilities with datasets proper inside Domo. It additionally allows Jupyter Notebooks to be scheduled as DataFlows to robotically refresh your information. By utilizing Python and Domo in tandem, KFC can now do the next:
|Import buyer opinions immediately from Apple and Android shops and mix them right into a single dataset||Schedule the Jupyter Pocket book to robotically refresh every day|
|Use Pure Language Processing fashions to establish the shopper’s emotion towards the app in every overview||Create a dataset that may be shared throughout the group|
|Extract essential metrics equivalent to when the overview was written and the consumer’s star-level score||Illustrate outcomes and metrics in a fascinating approach, utilizing firm branding and interactive visuals|
All of those options contribute to deriving insights for KFC’s cell app group. Now, the group can establish what works for patrons and what doesn’t, and domesticate concepts for future app enhancements—which all goes to indicate that when KFC clients converse, we hear. And that, in fact, is vital to long-term model and product success.