Friday, March 22, 2013

Big data & Intuit? Nahhhhh.. it will never work!

Big data & Intuit? Nahhhhh.. it will never work!


Intuit is extremely good at adapting to technology & changing priority based on the trend. However, for some genuine reasons (that I am not aware of) Intuit is lagging way behind in making use of Big data to enhance the customer experience, improve product offerings & generate more revenue.

What are the key challenges that I could think?
- Legal aspect of the data needs to be taken care of
- Adding ability for traditional products to send data/analyze data is a challenge
- Handling multiple streams of devices/technologies to talk to a single platform
- Willingness of the customers/user to share data is a big challenge

Options currently available to address the challenges
- Technical limitations to handle big data are not an open question anymore. With Intuit deciding to use Amazon Web services (Amazon cloud) to host web sites & manage web applications, we are all set to receive big data from customers
- With the introduction of smart phones, multiple streams of device can easily talk to each other today. We can sue simple proven technologies like SOAP or REST based web services to make sure Intuit data services are platform/device independent

Next steps...
There two big challenges that would require a mindset change & would require involving discussions. Legal aspect of the data to be collected can be handled by making sure we meet all regulatory laws/policies. Once this is taken care of, we have a clear segregation about what data can be collected & what cannot be.

Based on the above analysis, we can always enable the feature on a limited set of products to collect legally allowed data on cloud. Analyze the data collected & make recommendations that will enhance the user experience. We can always run a pilot product with limited number of customers to understand the way users are reacting to this. Users can be provided with an option to opt out from sending any data.




Examples on where we can make use of the data

Specific example 1: Vendor scenario or supplier scenario
A small business employer decides to set up a new door manufacturing business.
He buys quick books and sets up a new company file. When setting up a company file, the recommendation engine based on the location and the type of business selected sends recommendations about possible suppliers and vendors which will aid him to set a company with ease and trust.

For Example; door delivery business, door knob supplier, hinge supplier and/or supplier details can be recommended for this user.


Specific example 2: Bank example
In collaboration with financial institutions, Intuit can recommend other financial institutions  for the employer to save money.
For Example,  let us assume that the property insurance at bank ABC is being paid $1000 every year. If bank XYZ is providing the same insurance at $900 a year, by this recommendation, a user can use $100 each year. 
The same recommendation can be applicable for other expenses.

3 comments:

  1. Did you really mean: We can sue simple proven technologies like SOAP or REST? or use it? ;) And we really need to understand the legality of the data that is going to be collected. We cannot develop algorithms from the data without considering what data will be there finally after removing private ones. Good luck with your approaches..

    ReplyDelete
  2. on "Willingness of the customers/user to share data is a big challenge", I want to say that if you try to give user a valuable feedback on their data they will be easily agree to share on their data.

    For eg, on credit card data you can provide the feedback to user that on what they are spending and what they should do...

    ReplyDelete
  3. You have raised an excellent question, Manjunath. Intuit is a big advocate of "leadership by experimentation" (as opposed to leadership by powerpoint / politics) and hence one would expect that it will at least perform a few experiments before concluding whether adoption of big data is good or not. Your article provides good pointers for conducting such experiments.

    What would be great is if you can take a lead on a quick experiment that you can perform. All the best.

    ReplyDelete