Data Science is one of the hottest fields these days and has a multitude of business applications.
In this Data Science for beginners post, I want to give you an idea of what kind of questions data science can answer about your business. So you can set the right expectations and develop your understanding.
Generally speaking, data science can answer any business questions you might have, as long as the data is provided for the artificial intelligence model.
5 Categories of Questions that Data Science can answer for you:
1, Is it this or that? (A/B testing)
(Image courtesy: reliablesoft)
These constitute questions where you want to evaluate one condition vs another.
- Which will convert more customers: 10% discount or Free Shipping first 3 months?
- Which color bikini’s Gen X women will buy: single color vs multi-color?
- Did my conversion rate increase due to a particular influencer or new website design?
- Where should I post Free Shipping coupon, so users will convert more: Location A vs Location B?
- Will this marketing campaign fail in the next 3 months?
(Image courtesy: mubaris)
They include the category of questions when you want to cluster your users, customers, marketing campaigns etc in different categories to improve your understanding of your customers.
- Do I have a unique type of customers?
- i.e. create customer clusters w.r.t. location, age, gender etc.
- Do I have younger customers?
- Do I have mostly mature women over 50 years as my customers?
- Which combination of dresses in my fall collection can lead most purchase?
- What type of browser do my customers use?
- Which customers like the same type of shoes?
- Which days of the week does customer more likely to click on dresses vs shoes?
These are questions that can help you predict customer engagement, inventory etc.
- How much additional inventory should I order for Christmas this year?
- How many more red dresses should I order for Valentine’s day promotion?
- What is the likelihood that this customer will repurchase within the next 6 months?
- What is the likelihood that this customer will renew their subscription? (eg: if you run a monthly subscription service)
- What factors best predict demand for new Black Dress for my Fall collection?
- Should I change the store temperature to lower or higher or leave it the same?
- How many sales do I expect to see in India for this collection?
4, Are their anomalies?
This category of questions helps you answer if there are any anomalies and/or outliers in your data.
- If you see an increase/rise in sales on a random day where you didn’t run any promotion?
- eg: maybe Southwest ran special discount for tickets to Miami – that might result in increase bikini sales for you.
- Banks ATMs – looking for anomalies in ATM location usage.
- eg: if you use ATM in SF or bay areas. now all of sudden you are using ATM in Kansas, it might give the alert, of unusual behavior.
- Is this customer unsual?
- eg: if a customer buys 5X of product you typically sell.
5, Hypothesis Testing:
This category of questions helps you test if your hypothesis is true or false.
- Is there a relationship between time spent on jackets section & boots purchase?
- Will I lose this particular customer in the next 6 months?
- If I add testimonials to my product page, will it increase my sales?
to sum up …
Please note that data science is not limited to these 5 categories but data science can answer various business questions for you.
It will all depend on what kind of data you have been collecting on your customers and you can link it with the third-party data as well.
You can use this information for yourself to answer the question you might have, how to ask data science question.