Don’t worry, it’s not going to be one of those stories. People don’t usually talk about what actually happens behind the scenes of launching a system- a system like NextDrop. NextDrop is interesting because its a social enterprise, but unlike many other social enterprises, it does involve a fair amount of technology. And for all of you programmers out there, you know how nerve wracking it is to see whether your system has been successfully debugged, and whether it will work or not when you release it out to the public. Really, it’s incredibly stressful!
So we released our system out to the public yesterday. Now when you think of Silicon Valley tech startup launches, you think launch parties, media, and lots of free food. And I’m not really sure what you think of when a social enterprise launches (probably a blog post that says we have officially launched/something that documents success?) But let me tell you how NextDrop launched.
Expectation Managment: We didn’t actually tell them when the service would start which meant we had wiggle room. What I have noticed is that you need to figure out the least amount of information you need to give, but still keep people interested/still maintain credibility and usefulness. Key India Learning: If you promise something and you don’t deliver, you will be blacklisted by customers forever. But if you never promise it in the first place, people don’t seem to notice/mind. Finding that critical balance is the key, because then you know how much room for “error” you have.
Timeline of NextDrop “Launch”
Aug 30: We tested the system out with our own phones, and we were satisfied with the way the system worked. We were excited. We just had to wait for the water to come.
9:34 am: Get an admin NextDrop message saying Chandru has reported water on in Lingrajnagar North
9:38 am: Get an admin message saying that there were not enough residents to verify water was on
9:40 am: Check the call logs and find only 8 calls made (out of the 108 residents available to make to)
9:41 am: Realize that there is something wrong with the system
Conclusions: Swaroop, our tech genius, very quickly discovered the problem. We found out that most of the 108 people have placed themselves on the DND (Do Not Disturb) list, and since we run our service by making bulk calls to people, our calls were blocked to all but 8 numbers. And out of those 8 only 1 answered, which meant that the verification failed (NextDrop system needs a minimum of 2 calls to count as a verification).
Solution: Swaroop/Thejo (the person who built/designed the original system) are currently in talks with our calls/service provider to get us off the DND list (since we have signed consent forms from people). Until then, we have built in the manual push functionality- where Madhu will manually verify with people after receiving a valve open message from Chandru and manually push a valve open message to people (until we can get the issues sorted out)
Take 2: September 10th
5:15 am: 30 Minute Advance Notice for Patil Layout (since no verifications are required for advance notice, the messages are automatically sent to residents- i.e. our first success with messages going to 32 residents!)
5:16am: 30 minute Advance Notice for Lingrajnagar North (read: Success again with messages going to 103 residents!)
5:47 am: Admin valve open in Lingrajnagar North message received
5:48 am: Admin valve open in Patil Layout message received
5:49 am: Admin Advance notice message for Lingrajnagar South (successfully sent 78 residents the update)
5:50 am: Madhu verifies and sends out valve open messages to Lingrajnagar North, Patil Layout (Success!)
6:34 am: Admin Valve open in Lingrajnagar South message received
6:36 am: Madhu verifies and sends out valve open message to Lingrajnagar South (success!)
7:29 am: Call from Madhu saying that power went out so the water has stopped flowing
…and then it gets interesting
Now this may sound simple- we should just send everyone a “valve close” message and things would be fine. But its actually more complicated. We had discussed adding a third type of message which was the “last minute change in schedule” option for the valvemen to press for instances such as these. That way everyone knows that something happened and water has stopped flowing. Easy right? Here’s the thing we all need to realize. Information is a dangerous game- you have to know when it’s important to be as specific as possible, and when it’s not. Madhu brought this to my attention. He said that if people get this last minute change in schedule message, they need to know if it was a pipe break (which means they probably won’t get water today) or a power outage (which means they should wait a few hours because the water will come back on). This means that if we send out a message indicating a “last minute change in schedule”, everyone will call Chandru to find out whether its a pipe break or a power outage (read: ridiculous new burden on Chandru by having to answer 230 phone calls).
This is what is running through my head
The only way we can survive is if people trust us and trust our information. Here are the 3 scenarios that could play out:
Scenario 1: People got our advance notification, got the water from 7-7:30, and then realized that something unexpected to the water supply. However, they are very happy with the NextDrop service because it gave them advance warning of water and they collected at least 30 min of water before it went out
Scenario 2: People did not see the advance notice, and did not see the valve open message until after 7:30 am. So now they believe that the NextDrop service gives false information and they do not trust the service.
Scenario 3: They got the water, and collected it, but still think the NextDrop service is not useful because they didn’t tell them that the water would be turning off.
Management Dilemma: We don’t want to apologize if we don’t have to (i.e. raise expectations in case people fell into Scenario 1) but we definitely don’t want unsatisfied customers (i.e. we need to do something if most people fell under Scenario 2 or Scenario 3).
Action 1: Avinash, our business associate, called 10 people from Lingrajnagar North and 10 people from Lingrajnagar South (~10% of the customers in our 2 biggest areas) to quickly gauge customer reaction and to inform a decision
Action 2: I called Swaroop to find out how long it would take to build in the capability to let Chandru tell people that water has stopped because power is off.
Result 1: Avinash found out that all 20 people called had successfully collected water for an hour before the power went out and were incredibly happy with the NextDrop service.
Result 2: I found out from Swaroop that it would take a few hours to add the service/the ability for Chandru to inform residents of these changes
Conclusion: We decided to do nothing- less is more
..but it doesn’t end there
Things are going well, we are pretty satisfied until we get this message:
3:08 pm: Attigere Layout Valve Close (perfect!)
3:10 pm: Verified by residents that water is flowing in Attigere Layout (..uh oh)
3:11 pm: Attigere Layout: Water is flowing in your area (NO! Now residents have got another valve open message! This is wrong!)
So this is what happened: I realized Chandru calls the system right before he turns off the water, so when we call to verify that water is off, the water will still be on (because it happens instantaneously).
Action 1: Madhu called Chandru to ask him what to do (i.e apologize or send a valve close message)
Action 2: I called Swaroop and he realized what happens is that we don’t have a fail safe in our programming logic (i.e. we cannot send out a “valve open” message if the first message is a “valve close”). He quickly fixed the system, and was ready to send the people any message that was required
Result 1: Chandru told Madhu to just send a valve close message and things would be ok
Result 2: Swaroop fixed the system and pushed out a “valve close” message to the people in Attigere Layout. We also decide to put the “valve closed” updates on hold so it can be designed a little differently.
In conclusion, you can see what our NextDrop “Launch” looked like. It wasn’t perfect, and it was fraught with a lot of small fires to put out, but I think overall, the customers were happy and the NextDrop brand was in tact which is all that matters at the end of the day.
Also, another management lesson: Without an amazing team, this could not have been pulled off. Picking the right team is key, and yesterday made me realize that we have a really solid one here at NextDrop.
With a team like this, we can accomplish anything.