Below is letter I wrote to Sam Pitroda and Dilip Chenoy, both of whom kindly offered to help get a TRAI SMS exemption for Social Enterprises. I'd love to hear feedback from any readers too.
Jan 21, 2013
Dear Mr Pitroda and Mr Chenoy,
As social enterprises that strive serve India’s Base of the Pyramid populations economically and at scale, it is imperative that we can leverage SMS and IVR mediums to provide job alerts, financial services, produce alerts and more advanced services such job searches and interactive SMS services. Our users lack often the skills and money for mobile applications and/or web access and thus SMS and IVR systems are the only relevant digital mediums. The recent TRAI regulations have made many of these services virtually impossible to provide.
1. If we want to reliably send SMS messages to our users, they must be approved as Transactional by the VAS providers from whom we buy SMS bulk services. This implies that every message we send must be approved not only by our VAS providers but by the carriers from whom they buy SMSs in even larger quantities. The decision process to approve these messages is arbitrary at best and gets nearly impossible if our VAS providers have received any complaints from any of their other customers and cannot risk another complaint, less they get to a dreaded 4 complaints against them and lose their license to operate.
2. The decision to ban numerals in the SMS From field (implying all computer generated SMSs appear as from LM-SERVIC) have interactive SMS services impossible, given users cannot reply via SMS to any non-numeric From addresses. At Babajob, we have thousands of users who were searching our jobs by sending messages like “JOB DRIVER Mumbai” to 9972002222 who now cannot reply to any of the messages we send back to them. In essence, this move single handedly disenfranchised 1000s of informal sector workers who cannot afford smartphone and GPRS from being to do job searches via SMS, as well as countless other small, useful utility based SMS services.
We humbly request an exemption for social enterprises from the TRAI SMS and voice regulations (similar to the exemption other non-social services such as cleartrip, cab companies, all banks and Intuit’s txtWeb’s have already received) so that we can reliably deliver IVR and SMS solutions to our customers without getting every piece of content approved by VAS providers and carriers.
We would be happy to discuss this request and brainstorm alternative solutions or frameworks to address this problem too if you desire.
We look forward to hearing from you,
Mobile: 91 98886 251476
We recommend that every Social Enterprise sign off on a code of conduct whereby the services are used to provide their own relevant services.
Every Social Enterprise would be required to:
1. Keep electronic records of all sign-ups, opt-ins and invitations of their users with mobile numbers for audit purposes
2. Register any ID or mobile number they use for exempted interaction with TRAI.
3. Provide the VAS Provider or SMS Gateway that provides SMS/Voice services for the Social Enterprise, and update this information if it changes.
4. Sign off on the Social Enterprise SMS/Automated Call Code of Conduct.
This Code of Conduct would include:
• Only provide or advertise services that are directly relevant to their mission as a Social Enterprise.
• Send SMSs and Calls only to those users that have interacted with the Social Enterprise in the past or those users who were recommended or invited by existing users of the Social Enterprise.
• If the Enterprise sends more than a total 500,000 SMS/Calls or sends regular (more than 10) SMS/Calls to users, the Enterprise will inform users users how to STOP or opt-out the service and how to file a complaint against the Enterprise.
• Keep their numbers and ID to a reasonable minimum, so that users generally have a well-defined number with which they can interact and file complaints.
Failure to comply with the Social Enterprise Code of Conduct would cause the Enterprise to lose their exemption and thus be treated similar to any other VAS service. Determination of failure compliance would rest with TRAI.
The Exemption would grant the following rights to the Enterprise:
• The ability to send SMSs and Calls to all their users, including those on DND lists
• The ability to provision two-way interactive SMS numbers, where by users can send an SMS, call or missed call to a the Enterprise’s number and receive an SMS or Call back with the same numeric number, to which the user can directly reply.
The exemption cannot be transferred to any other organization.
Any organization can apply to TRAI as a Social Enterprise, assuming they provide socially relevant services, submit to the Social Enterprise Code of Conduct and are a registered Indian company or NGO. Supporting letters from Social Enterprise industry associations or other references may also be provided.
We can imagine the following objections and have provided a summary of our potential responses:
1. How do we ensure that the exemption is not abused with more unsolicited calls and sms?
a. See the Code of Conduct above. If an Enterprise breaks this code, TRAI can remove their exemption
2. How to prevent calls and sms from people when roaming
a. Given that many users do travel, these often are vital parts of the experience that an Social Enterprise wants to provide. That said, the Code of Conduct makes provisions for an Enterprise to inform users how to stop the service.
3. For security reasons need to limit number and also register persons sending sms
a. As stated above, all enterprises must register their numbers, their providing VAS Provider and make efforts to keep their number use to a minimum.
4. Who determines what is a social enterprise and thus, eligible for the exemption?
a. No one explicitly makes this determination. Rather all Enterprises must submit to the Code of Conduct.
I met an interesting fellow yesterday. Young, brash, self-confident. Just quit a NYC consulting job a year out of Stanford grad school to figure out his next gig.
Speaking with him was a bit like looking back at myself just as I quit Microsoft to start Babajob.com in 2007. In particular, I was struck by how he did not know what he didn’t know and had not yet learned humbleness in attempting to solve big problems, a sentiment that smacks of getting old. “I am successful and I will be successful in doing this even if I fail,” he uttered and was a quote that I said to myself almost verbatim when I chose to leave corporate life. It’s in all likelihood true for both of us, but it does not guarantee the success of the enterprise, just that we would both be able to find another job if we failed. But a personal safety net is not a plan to change the world, even if it makes us feel more like we can.
This brashness and naivety may in fact be essential to an entrepreneur – the suspension of disbelief – especially in one’s own ability, the feeling that all problems are solvable, and perhaps too the ignorance of things that are inherent risks to the enterprise. These may all be necessary elements of doing an activity as risky as a startup.
After 6 years of thinking about Babajob and working in this now-VC funded but still just a 30 person startup, I think can I distill which of those youthful elements I respect and which are not recommended.
First of all, a new enterprise is an exercise in hope. A founder needs a profound hope and trust in himself or herself and that hope and belief must be infectious to attract employees and investors. It also makes work fun and is vital when things inevitably get hard. If hope and confidence looks a bit like arrogance, that’s OK.
Second, a vision of change regarding the new picture that the founder wants to the world to be if their enterprise is successful is also needed. Important companies need big ideas – A desktop in every home, the world’s information organized and available at every person’s fingertips, and of course, a access to better jobs – even for the poorest – available at the touch of the mobile phone. This vision of change is deeply connected to the hope that the founder needs to instill in their employees and investors but also makes the startup press-worthy and noteworthy to wider social circles, which is how the founder attracts potential employees and investors.
Third, ignorance is a sticky one for me. A founder can’t know everything about their field when they start, but a deep articulation of the risks and unknowns is something important and an element I wish I had spent more time thinking through earlier. It’s an element that I frankly rarely see in young potential entrepreneurs; they have a reasonable idea of the problem they want to solve and (often too much) detail about the solution they want to provide but often lack sufficient insights around the ways their solution may fail and how they would evolve it in the face of that failure aka new market insights. This last element is of course the hardest to predict and this evolution of failure and iteration is the stuff of actually running a startup.
That said, as our own investor Vinod Khosla and others like Eric Ries have pointed out, a startup is an exercise in doing something highly risky and the systematic reduction of risk is actually the point of the startup. Hence, an articulation of the important risks and a plan to assess them quickly could literally save years of iteration time.
There’s a tension here obviously– namely that if the founder truly appreciated all the risks of failure of the endeavor, she might never actually the company.
Still, wide-eyed hope seems preferable to blind optimism. Choose problems with risks you understand or at least are aware where the hidden risks may lie and your odds of actually changing the world get much better.
Assumed Goals of the State (say Karnataka):
1. Attract the best and brightest Indian start-ups and funders to target Karnataka as a market and place to do business
2. Encourage local-language applications that provide societal value to a broad audience
3. Encourage new business models to thrive - e.g. Flipkart is not dependent on the telcos to collect money
4. Maximize the utility and usage of services, to both create value for citizens and create new potential tax revenue sources in the long term
Try to choose winning technologies, companies or business models
Thoughts on How:
Reduce rather than increase regulation - the recent TRAI regulations killed a few of the companies like MYToday that did not have a telco dependence, led to avalanche of service disruptions and exemptions for well-connected companies. This has scared off any remnant of investor capital in Mobile Value Added Services, pushing it to the largely unregulated ecommerce market (and making it even harder for companies that wish to provide information services to the masses over the phone). This clearly is an awful policy.
Encourage new technologies - the focus on regulating SMS and UssD is simply pushing people to smartphones, which frankly are far more capable to deliver rich, local-language, offline capable experiences than the dumb phones most targeted by most traditional MVAS companies.
Solve real, hard problems for new companies:
* Cash collection- enable at service to use Bangalore One kiosk for cash payments at a reasonable fee.
* Make “Know Your Customer” requirements and identity verification easier - UID, etc
Encourage cheaper citizen access to services - the bonanza of spectrum prices for government coffers has given India 3G networks that are absolutely more than the Common Man can afford. We need lower tariffs for 3G as well as free Wifi locations across the state.
Absolutely, do not proselytize specific technology solutions or platforms. I can think of no government technology platforms worldwide that are a hotbed of innovation compared to the Apple AppStore or Google Marketplace. The primary success metric must be customer adoption and unfortunately, these government technology platforms tend to have abysmal records in gaining such adoption.
Let the government - and its tech providers - provide API access to data and services - not the customer experiences. Encourage small players with new, untested business models and do not charge companies for usage of APIs until a given company makes over say 1,000,000 API requests per month.
Provide transparent incentives to mid to large websites that create mobile apps experiences in Kannada. E.g. If you have more than 20,000 monthly visitors in Karnataka according to Google analytics, you will receive Rs 10,000 if you translate your site into Kannada.
Provide the best APIs in the country and leverage this as a unique advantage of working and creating apps in Karnataka. -The state government already has access to the location in real-time of every phone for security purpose; what if a hackathon could be created from this anonymzed data to build better traffic patterns and study daily migrations?
Like a real start-up, start small, take steps to reduce risk BEFORE each successive investment, validate assumptions of citizen behavior and usage before expanding program, etc.
I've been thinking about this space a fair amount with Freeman Murray @ http://jaaga.in
and here are my initial thoughts
We will soon have the platforms – phones, cheap tablets and reasonable 3G data connections – to enable a potential revolution in the Indian educational system. Perhaps, but lots of folks said TVs in the classroom would make education significantly better and it never happened. That said, given the interactive, ubiquitous and personal nature of cheap tablets and touch phones, this surely has greater potential.
Elements required in the solution:
1. Cheap Data – right now one 500MB Bollywood movie downloaded over 3G costs Rs 500 - $10 – aka 5x more than seeing it in a theatre. And 10x more than what the bottom 50% of India pay for their entire monthly mobile phone bill. It is frankly appalling that the government taxes 3G licenses to the tune of $40bn and then asks for cheap bandwidth. We need innovative telco incentives and business models that the poor connect to connect via their normal mobile carrier connections.
2. Localized Content – Just because India will have a great mobile and tablet ecosystem, does not mean that English language literacy changed overnight. The Khan Academy is great if you a smart 17 year-old who speaks fluent English but few 10 year olds in rural villages are going to understand Mr Khan’s American accent. We need content that’s taught in local dialects by great local teachers on locally relevant subjects.
3. Community and Face-to-Face learning – Face it – watching a bunch of educational videos by yourself if not fun and most kids don’t learn in asocial environments. We need to take a page from the Digital Study Hall and DigitalGreen.org – use the technology to create occasions where local groups connect and learn together in real life motivate each other.
4. Student Recognition and Incentives – Kids will need incentives to watch and learn from educational content rather than watching Bollywood movies (which will always be just one app tap away). We need to create ways that students who do demonstrate skills acquisition and learning from digital platforms are recognized by their peers, the media and in their communities (just as toppers of the college entrance exams are today)
5. School and Teacher engagement – given that most learning is social, the most logical place to accelerate learning is in classroom. Again taking a lesson from Digital Green, teachers need to be encouraged to create best-of-breed digital lesson videos and interactive modules on tablets and compete to get those seen and distributed in their local communities. We need “Teacher Idol” replicated in 100 regions across the country and viewed on a 100 million tablets so that exceptional teachers across India get the recognition they deserve.
6. Educational Research and Government Support - The business models for how one makes money creating great educational content communities for the poor in India are yet clear, but they certainly have the potential for massive and scalable society benefit. Government and philanthropy need to provide research money to try out lots of ideas. Some may yield profitable business models but given that public education is a good in its owwn right, the focus must be on improving educational outcomes at scale, not necessarily making money immediately by selling to wealthier students.
Here's the video!http://www.youtube.com/watch?v=sYlC0GtKpHg
I'm giving a talk at TEDxNIDBangalore http://tedxnidb.com
on October 1 and I've been debating what to discuss.
Topic 1: Societal Trust and Evolution vs UID and the End of Privacy.
In short, my views on how evolution and terrorism predict greater levels of trust and verification in society even though technology is making this terrifying. Imagine yourself with a Bangalore beat-cop in 4 years and he asks you why one of your facebook friends made a call to the cousin of a Pakistani terrorist while you were travelling near Leh in December, 2013?
Topic 2: Efficient Markets for the Poor
How will services like flipkart, babajob and justdial benefit the bottom 80% of India in 4 years? Is everyone gonna have apps? What effect can we expect this to have on normal people and businesses in the years ahead?
1. Expose the location data (it’s already available anyway)
2. Expose the social graph
3. Provide open, uniform billing for non-phone SIMs
4. Expose the retail and digital payment system
5. Expose your verified profile data
In general, telco folks don’t think about their assets the way Google, Apple and Microsoft think about theirs. Namely, how do I leverage my assets and offer them to other businesses that bet their livelihoods wholly on top of the hard problems I’ve already solved. To name some examples, Apple solved the micro-billing, device compatibility, application-update and consumer trust problem for application developers with uniform rev-share billing APIs, a limited number of handsets, background updates and rating systems available from launch, respectively. No carrier has arguably solved any of these problems for their mobile developers. Instead in India, they have created a market for non-SMS VAS applications that primarily have little utility and frankly hover at 4% of revenue only with excessive promotion and tricky subscription billing schemes that users often accidently turn on and are difficult to turn off.
It’s not just an Indian problem though; by not focusing on providing platforms that enable developers to write interesting applications on top of them, telcos have essentially made the choice of carrier into a dump-pipe, commodity game – which network has the cheapest, most reliable voice and data service – rather than which network has the best apps?
But telcos have solved many of the hardest problems facing smaller businesses today and I’ll argue that enabling the telco as a platform is the key to unlocking their next wave of revenue for phone companies.
1. Expose the location data (it’s already available anyway)
Telcos have always known where your mobile is - how else would they be able to route a call to the right red and white eyesore (aka cell tower) on your way to work? Mobile location-based scenarios have been discussed for 15 years but it’s only in the last 2 years that these applications have seen any significant user adoption (e.g. google maps on smartphones, foursquare). What’s ironic is this generation of location-based apps entirely ignore (via GPS) or reverse engineer the telco’s location data by measuring the signal strength of nearby towers and then look up where the towers are in databases maintained by Google, Microsoft, Navtaq and others. In other words, if your phone has a data connection, any application on it can roughly determine where it is without the telco ever being the wiser. Thus, the location data that telcos always sought to charge app developers for is now free. Furthermore, other business models have developed around local apps that the telcos will never share in (imagine paying your phone $.50 per foursquare check-in).
But ultimately, these methods are work-arounds and limited. GPS and cell-tower triangulation does not tell an app developer anything about where all the phones are in a general neighborhood, what the car and traffic patterns are or how many people actually attended a given concert or movie– all things the telco actually knows in its data centers but only exposes when the bomb-squad comes knocking.
Thus, my free advice. Provide a set of free, secure, cloud-based, use-limited APIs that aggregate where all the phones are and APIs that streams a given user’s location (with their permission of course). If an app developer makes more than 100 million reads, you know they have a business model and then start negotiating how much you should be paid.
2. Expose the social graph
It amazes me that the telcos have always had the most interesting social network graph – the people I call – but it took a geeky twenty-something to turn that social graph data into a bazillion dollar business called Facebook and then enable “viral spread” to spawn other billion dollar businesses like Zynga. Clearly, the telcos missed the boat on this one.
But all is not lost for our friends with spectrum. Combined with location data and verified contact data (see below), the social graph created by our call patterns arguably still represents the most accurate representation of who is important to me. Obviously, app developers should be able to leverage this data as easily as they can call the FBAPI or Google’s opensocial.
3. Provide open, uniform billing for non-phone SIMs
We’ve reached the stage where it makes sense for lots of devices to have wireless data connections e.g. Kindles, iPads, cars, my keys, my dog’s collar or any object of value (so I can track and find them). It makes less sense that consumers have to pay a monthly data subscription plan each of these devices, especially if the amount of data they use is small and their utility is narrow. Unfortunately, there’s no easy way that a small developer or device maker can simply buy a small set of SIMs – say for a $7 a piece – that are authorized to send or receive 1GB of data and have a lifetime of 5 years. As a small developer, I should not have to negotiate a deal with someone inside a telco’s biz dev team (as Amazon has with the Kindle), to build a few thousand devices with a wide-area data connection. Give the developer community simple, ease-to-use data pricing for devices that can be applied to any SIM, and you’ll see that wirelessly connected devices will explode on a per capita basis.
4. Expose the retail and digital payment system
I like to point out that the American Internet is the cheapest place to buy anything on earth. There are lots of reasons for this but one of biggest contributors is ubiquitous digital payment instruments – credit cards, paypal, etc – along with cheap shipping and low tariffs across states. In India though, if I make a physical or digital good, I can only really sell it broadly if I have the ability to collect cash from my customers. This is hugely annoying to a digital company such as Babajob.com or Cleartrip, given that less than 3% of India has a credit card. Flipkart is arguably solving this problem beautifully by sending a dude on a bike to my house to collect my cash payment when they deliver a $3/Rs 150 book, but really – does every online company want to re-create its own nation-wide cash collection network of guys on motorbikes?
Now, the telcos arguably have the best network of cash collectors in the country – literally 2 million tiny shop owners make their livelihoods earning Re1($.02) on Rs 50 ($1) mobile recharges. Of course, it’s difficult to organize but there’s no reason they could not turn 50,000 of those guys into the 7-11s of Japan - where products get delivered to a nearby corner store and consumers walk down and pay a fee to pick up the product. Especially for digital products like services or tickets, it’s easy to imagine a simple system where any consumer can give their recharge shop Rs 300/$6 for a product (with a known unique number) and the telco takes 5-7% of the transaction (which is what most Indian credit card processors charge). Just as early mail order companies drove their customers to use MasterCard and Visa and ebay drove users to paypal, 100,000s of businesses would drive their customers to particular telco recharge stations if they could leverage them to collect cash at reasonable rates (rather than the 80% rates that VAS companies live with today).
5. Expose your verified profile data
The next five years in India will see a hugely important rise in verified user data. Very soon, in order to get a SIM, you’ll not only have to give proof of your address and a government ID (as you need to provide today), but also a biometric UID-mandated finger-print and eye scan. This data collection will be mandated by the government (because it’s really handy in tracking down terrorists) but the data is incredibly useful for anything that requires trust among people. This includes job profiles (where I’m absolutely positive we at Babajob.com can help job seekers earn more from employers if the employers know they are verified individuals with known addresses) but also enables a host of other vitally important trust-based services like credit-agencies, loans to individuals (rather than self-help groups), etc. There are few initiatives that could transform the economic potential of India as greatly as better trust and verification systems. If the telcos can state their users are verified, UID citizens, capable of signing contracts (where default suddenly gets much harder), it massively increases the value of their customer database for almost any business.
The first Indian telco that creates a scalable, simple, near-free verification and profile data reuse API is the telco that gets to be the backbone behind conceivably billions of transactions per day in the next decade.
That’s it for now. I hope the telcos are listening; they are among the most important enabling institutions in our societies and its high-time they started acting like it.
I had a nice chat with a delegation from Tajikistan (population: 7mil) this morning who represent the advisors to the President and Parliament on mobile and Internet policy.
They liked babajob but the meeting got me thinking about what policies I'd put in place if I were in their shoes.
1. Encourage mobile digital payment systems. Google - and all free internet services - only exists because ad publishers pay for clicks. Publishers only pay for clicks because a reasonable portion of those clicks turn into actual sales on their website. Those website sales only happen because people can buy things online with their credit cards. It's the reason that Google only made $20m in 2008 in India (even though it had the #1, #2 and #4 sites) - we have only ~20m credit cards (and very few people use them). You don't build digital payment systems because you want to make Visa rich - we do it so that an entire ecosystem of new services can arise.
2. Similarly, policy makers should be encouraging mobile banking systems far more quickly than they do today. These are two of the most heavily regulated industries in the world but they really should merge. If you've visited the Philippines or Kenya anytime in the last 5 years, you'll see the amazing power that being able to super-easily move money on phones has enabled. It's really only because of regulatory hurdles that these systems have not been copied elsewhere.
3. This is more specific to India but governments must issue unique identifiers to their citizens (e.g. UID, social security numbers, etc) in order for credit systems to properly function - be they MFIs that get duped because one woman belongs to 2 loan groups or credit agencies that do not exist India because they can't correlate their user records. Yes their are privacy concerns in UID but read Jared Diamond or Robert Wright - the societies that have thrived over time are those with strong trust, land title and contract systems. In a modern society those systems are entirely based on knowing the difference between two people with the same name, something that's made possible if you give everyone a unique unumber.
4. Encourage efficient markets and communication for all citizens- even the very poor. This is somewhat a plug for babajob but the poor do face the most inefficient markets in the world and that often fundamentally limits the choices (How are you supposed to look for a better paying job if you are a live-in maid in the country? Well, we are trying to solve that problem). Furthermore, super-cheap cell phone plans improve the quality of life and allow for much better real-time coordination of activity. For example, the real reason that rural fisherman with phones make money is not just that they know which market has the best price - it's that they can call other fisherman who tell them where the fish are today. Similar, if a delivery helper in Mumbai has a phone, it's easy to imagine that the number of deliveries he can make radically increases because he does not have to wait around (he can just call ahead).
4a. There are 2 primary ways that governments can ensure that phone calls are cheap:
1. Keep the regulated rate that carriers charge each other for connecting between networks as low as possible. This is the reason that India has some of the lowest rates in the world (and the US's are so high). Of course this rate is guaranteed profit floor for the carriers and they'll resist it mighty, but this is way to make the market bigger and efficient.
2. Don't overcharge for 3G bandwidth. I honestly find the 3G auctions ridiculous. Rather than taxing the potential innovations of 3G AFTER we've found the killer applications and a market has been proven, governments are taxing the carriers before BEFORE any significant money has been made. This effectively makes cheap 3G systems impossible (because the carriers had to pay billions for the spectrum). It's really the exact opposite of the policy you'd want to encourage innovation. Instead, just levy a tax on the services when they are consumed.