Here is my take on the problem statements which were of particular interest to me. For ease of scanning, these have been structured as follows:
Increasing customer lifetime value and retention for Licious' on demand meat and meat based products
Increase the number of returning users who have purchased previously bought items or high frequency items on a regular basis.
Increase number of repeat purchases
Increase number of returning users
Decrease in drop off in payment funnel for returning users
Understanding user opportunities
From the user’s perspective, it is desired to decrease any friction that is caused due to multiple checkouts and reduce the cognitive load by making sure the user doesn’t have to remember to buy an item everyday at a particular hour.
Understanding business opportunities
From a business’ perspective, it is desired to increase the customer lifetime value and in the process give a larger window to provide value to the users. This would mean increased engagement with Licious as a service and not necessarily with the physical app. As they will be trying to subtly change their purchase habit, it would help the retention figures to improve.
Daily Subscription Model
A daily subscription model in e-commerce is one where users can pay upfront for the next “n” days with “x” days of gaps in between and get products delivered at their preferred time slots. This is a one time process where users need not purchase those items that are purchased on a regular basis and rather have to perform the checkout procedure once. This is needed especially for those users who frequently need to purchase the same item (like milk) on a regular basis to avoid the repeated hassle of ordering items.
Purchase frequently bought items with my least intervention.
Hassle free payment of items that are purchased on a daily basis.
Improving customer experience and establishing trust by designing Ninjacart's first version of the Dynamic Serviceability Platform
To arrive at a decision regarding the acceptance of the order based on its serviceability, which is based on a few real-time parameters.
Percentage of users served on time
Percentage of cancellations when the ETA is exceeded
Ratio of orders delivered on time per geo-location to total orders delivered
Measuring the size and urgency of the problem
Let’s say there are roughly n stores in the order of 100s and m products in the order of 1000s per store, which are in the vicinity of a customer location. This would mean that for every request to display a product made on the app, we will need to find the distance from these n+ stores to the location. Let’s assume DAU = 100k users with approx. 100 users sending a request/ min. # of requests per minute =100*m*n
Say requests can be made in any of the three pages - homepage, search, category, add to cart and place order; then there are 5*100*m*n requests per minute= 5*100*1000*100 (MN)=5(MN) million requests per minute!
Why build this feature?
To avoid delivery related problems that occur during both the pre order phase and the post order phase- here we are concerned about the pre order phase as it is crucial to arrive at a decision regarding the acceptance of order at this stage before proceeding to the next phase.
Users - To establish trust between the users and the service/s promised in the pre-order stage.
Business - To increase retention and referral driven acquisition as customers will return to the platform and word of mouth will increase if services promised to them are fulfilled or exceeds expectations.
Dynamic Serviceability Platform - General
Determine which stores/products to display on customer app based on the customer location input.
Determine the maximum distance between a store and the customer location for which a delivery can be made/ is feasible.
Determine the factors on which the maximum distance will vary (according to the geography and operating environment )
Find optimal route from store to customer location.
Determining surge pricing that is to be charged to the customer
Provision for checking capacity of the delivery fleet corresponding to the particular network of stores.
Dynamic Serviceability Platform - Delivery Agent
Determine stress on the delivery fleet.
Predicting time to deliver an order depending on a. Delivery agent assignment time, b. First mile time, c. Preparation time, d. Wait time, e. last mile time +distance
Improving upon an existing job discovery and job application platform for the blue and grey collar workers and recruiters
Number of job offers received by a user / Number of applications sent
Number of users who received a job offer / Number of applications sent by the user
Average time taken to fill up a vacant position
Why is it important to the users and the business?
For the stakeholders(users), the delay in the hiring and shortlisting process often results in candidates losing interest in that position and recruiters losing out on potential candidates. For the business, it is important to maintain the user’s satisfaction and meet their needs without which the number of active users will start to decrease.
How did I measure the size + urgency of the problem?
According to Indeed’s report, the reasons for ghosting among job seekers varies, with 20% saying they received a better job offer in the meantime, 13% dissatisfied with the salary offered and 15% indicating they decided the job wasn’t the right fit for them. This is a major problem that needs to be addressed otherwise this ghosting trend is going to be on the rise leading to user dissatisfaction.
Mandatory prioritising of applications by assigning a rank against each job application
Case 1 : From applicant’s side, Step 1- Apply to a bunch of jobs at the same time. Step 2- Rank them Step 3- Make edits of ranking before the application is seen and reviewed.
From recruiter’s end, Step 1- Get the list of candidates who ranked their applications as 1. Step 2- Only if the requirements are not met in the first list, he will move to the second set of list of candidates.
Case 2 : From applicant’s side, Step 1- Apply to jobs on a daily basis. Step 2- Rank these applications as and when applied. Now, edits in rankings are allowed before the recruiter reviews the application.
Grouping applications into similar cohorts or groups
Step 1- Find similarity among profiles and group applications received into different segments. Step 2- Find a common denominator in each group that has the highest similarity percentage within that group.
Step 3- Rank profiles of each group in descending order of percentages.
Step 4- Recruiters will have to make a fictional persona with traits set as per requirement and then shortlisted profiles will be vetted and the most close match will be selected.
Why doesn't Uber connect with my Google Calendar to be aware of my upcoming events & nudge me to book a cab basis availability, traffic etc so I can reach my destination on time ?
From a user's perspective, the underlying problem is "reaching the destination on time" where the stated assumption is "an already defined travel and the unstated assumption is "uses google calendar & prefers to use ride hailing services". If reaching destination on time is the only problem here, we just need an alert/reminder in the form of a push notification etc from that respective site from where we booked tickets for movie/flight/etc. It would be a business opportunity for those sites. I fail to understand why Google Calendar needs to be connected specifically with these ride hailing services when GC already gives a reminder some x minutes before the event. Unless of course we want to solve the problem where we underestimate the time it takes to book a cab, GC handles it pretty well. But this doesn't solve our problem of users reaching destination on time. What about users who haven't marked the event on calendar (as many have pointed out) ?
Here comes the business opportunity for ride hailing services like Ola, Uber . 1. Google calendar integration Pros - It does solve the 2 problems mentioned above. Cons- Not many people use it/record the event. It is more of an opportunity for GC
Verdict - 👎
I thought about Ola/Uber integrating with these sites/apps and fetch the date and time whenever a user booked. Well, that would mean partnering up with tens of websites and apps. Tedious!
What is the common event/process that a user performs while booking tickets? Payment transactions!Assuming we are considering online payment transactions using UPI why not integrate that API with the ride hailing ones? Let these UPI apps fetch data. Kidding? No.
Scenario 1 - user books tickets via app/site --> makes payment using UPI--> UPI sends data (time and date?) to Uber/Ola--> Uber/Ola sends notif/prompts to user to book/pre schedule a ride.
Scenario 2 - user books movie/flight tickets via UPI apps like Google Pay , PhonePe etc--> UPI sends data to Uber/Ola --> Uber/Ola sends notif/prompts yo user to book/schedule a ride.
Location detection on adding address
Licious vs Zomato
Licious - Born to meat -
Click on "Add address".
Automatically input locality/area.
Manually enter flat no, building, street no etc.
Identify location based on input.
If wrong location is detected, move map to change
Search for location OR use current location
Pin dropped at exact location OR move the pin manually on the map.
Show number of valets near a restaurant
What do you all think is the point of knowing the "number" of valets near a restaurant when ordering food through Zomato(say)? Even if there's no valet and we anticipate that it's going to take more time than we can afford, we can't cancel the order. But, that's beside the point. What is interesting is why they tweaked the previous feature of showing the real time "location" of valets around the restaurant.
Now that I think deeper, Ola and Uber have also followed a similar revision of this particular feature.
Pause watch history and search history
Pause watch history and search history features in youtube are a great addition over the conventional delete all watch and search history.
Useful in situation where a colleague or a friend uses the user's personal account to watch videos as recommendations depend on user history and the user won't want other's history to be used for recommendations.
Pinterest has tabs on the top bar where some of the frequently used customised pin boards created by the user are used to generate more such content and shown as a feed.
Quite exciting to see the thought behind this- users are more likely to consume similar content based on their interest in a structured way.
Quite an addition to the more common filter option.
So this feature along with a generalised feed section showing all recommended feed haphazardly is a welcome addition
Samsung photo gallery
Samsung photo gallery application has a search feature where the documents category has sub categories - notes, presentation files, maps, newspapers, etc that is otherwise not present in the more popular google photos.
Interesting to see the accuracy of the feature deployed using AI
Two adjacent virtual back keys
The functionality was that one would function as a normal "return to previous page" key while the other would redirect to the home page/landing page without any intermediate stops(yeah! A non stop flight saving a lot of "user taps" aka fuel).
Does this have something to do with the loading time of the pages? Or is it to eliminate false positives/redundancy of clicks(when only the total # of clicks of return button is captured and hence not paint the full picture)?
Don't you think that if someone has to go the long way to return to the home page and has no option to skip the intermediate steps, there is a high possibility that the user might get hooked to a particular product/event and eventually perform a crucial action(like adding to cart or purchasing)?