Wednesday, October 31, 2012

Building And Expanding Enterprise Software Business In Brazil

While in Brazil, describing his country, one of my friends said, "We have all the natural resources that we need to be a self-sufficient country and we have had no natural disasters such as earthquakes and hurricanes. The only disaster that we had: we lost the worldcup."

This pretty much summarizes Brazil.

A helipad in front of my hotel in São Paulo.
On one side, São Paulo, the seventh largest city in the world, has one the largest per capita helipads in the world where the rich people don't like to drive around in traffic in cheap cars to avoid getting kidnapped at stop lights. On the other side, it is one hell of a city, just like Mumbai - large, organized chaos, and money. It is growing and it's growing fast. While income inequality has been on a steep rise in emerging economies as well as in the western world, it is declining in the Latin American countries, especially in Brazil.

If you're thinking of building or expanding enterprise software business in Brazil, now is the time. This is why:

Developing to a developed economy

Brazil has been boxed into BRIC economies but in reality it behaves more like a developed economy with lingering effects of a developing economy. Even though corruption is rampant in Brazil, it exists at much higher level and a common man typically doesn't suffer as miserably as he/she would suffer in other countries such as India. Being a resourceful country, there are all kinds of jobs. The bureaucracy will break and the infrastructure will also catch up very soon due to the soccer world cup in 2014 followed by the Olympics in 2016. Don't apply your BRIC strategy to Brazil. Consider Brazil as a developed nation and aggressively expand.

Courtesy: Economist
Stronger middleclass

Middleclass has money and they are willing to spend. Brazilian tax laws are the most complex laws that I have ever seen.  Even though the global retail brands are present in Brazil, they are outrageously expensive. Making a weekend trip to Miami to shop is quite common. Even after paying for a plane ticket and hotels it is cheaper shop in the US. The retailers in Brazil are trying to better understand this behavior and the global brands are also looking at several different ways to market to this middleclass. As an ISV this is a gold mine that you should not be ignoring.

Local to Global

Following the nation's growth many local companies in Brazil are aspiring to go global, establishing their business in developed economies. Local ISVs neither have scale nor features to support these efforts. These companies (typically mid to large) are looking at global ISVs for help, and yes, they are willing to spend.

Then you ask, if it is this obvious, why aren't global ISVs already doing this. They are. It's obvious, but it is not that easy. These are the challenges you would run into:

Complex localization

Many global ISVs have given up localizing their software for the Brazilian market. The tax laws are extremely complex and so are other processes. If you are truly interested in the Brazilian market you need to build from scratch in Brazil for Brazil. Hire local talent, empower them, and educate them on your global perspective. Linux and related open source software talent is plentiful in São Paulo. These developers are also excited about the cloud are are building some amazing stuff. I would also suggest to either hire or partner with local domain experts as consultants, who can work with a product manager, to truly understand the nuts and bolts of local processes, laws, and regulations.

Rough sales cycles

Selling into large accounts is not easy. Work with partners for a joint go-to-market solution or have them lead or participate in your sales cycle. The sales cycle is not fair and square and the purchase decisions are not just based on merits of your offering. Even if customer likes a product, commercial discussion are a huge drag, from the sponsor, to buyer, to all the way up to purchasing. Be patient and take help of local experts to navigate these roads.

My taxi driver watching a live soccer game while driving
Language and culture

Speaking Portuguese is pretty much a requirement to get anything done. But, if you speak Spanish, you could get around and also pick up a little bit of conversational Portuguese. English-only approach won't work. Do not even attempt. Also, Brazilians don't like to be called Latin Americans. They like to be called Brazilians; avoid any Latin American references. While you are there, learn a thing or two from an average Brazilian about fitness. Unlike Americans, the Brazilians are not into junk food. At a churrascaria, they eat salad followed by meat followed by some more meat. If you wonder why they are so fit, especially in Rio, this diet perhaps explains. They do enjoy their lives and sip Cachaça at the beach, but they are damn serious about working out.

Tuesday, October 16, 2012

Analytics-first Enterprise Applications


This is the story of Tim Zimmer who has been working as a technician for one of the large appliance store chains. His job is to attend service calls for washers and dryers. He has seen a lot in his life; a lot has changed but a few things have stayed the same.

The 80's saw a rise of homegrown IT systems and 90's was the decade of standardized backend automation where a few large vendors as well as quite a few small vendors built and sold solutions to automate a whole bunch of backend processes. Tim experienced this firsthand. He started getting printed invoices that he could hand out to his customers. He also heard his buddies in finance talking about a week-long training class to learn "computers" and some tools to make journal entries. Tim's life didn't change much. He would still get a list of customers handed out to him in the morning. He would go visit them. He would turn-in a part-request form manually for the parts he didn't carry in his truck and life went on. Not knowing what might be a better way to work Tim always knew there must be a better way. Automation did help the companies run their business faster and helped increased their revenue and margins but the lives of their employees such as Tim didn't change much.

Mid to late 90's saw the rise of CRM and Self-Service HCM where vendors started referring to "resources" as "capital" without really changing the fundamental design of their products. Tim heard about some sales guys entering information into such systems after they had talked to their customers. They didn't quite like the system, but their supervisors and their supervisors' supervisors had asked them to do so. Tim thought somehow the company must benefit out of this but he didn't see his buddies' lives get any better. He did receive a rugged laptop to enter information about his tickets and resolutions. The tool still required him to enter a lot of data, screen by screen. He didn't really like the tool and the tool didn't make him any better or smarter, but he had no other choice but to use it.

Tim heard that the management gets weekly reports of all the service calls that he makes. He was told that the parts department uses this information to create a "part bucket" for each region. He thought it doesn't make any sense - by the time the management receives the part information, analyzes it, and gives me parts, I'm already on a few calls where I am running out of parts that I need. He also received an email from "Center of Excellence" (he couldn't tell what it is, but guessed, "must be those IT guys") whether he would like to receive some reports. He inquired. The lead time for what he thought was a simple report, once he submits a request, was 8-10 weeks and that "project" would require three levels of approval. He saw no value in it and decided not to pursue. While watching a football game, over beer, his buddy in IT told him that the "management" has bought very expensive software to run these reports and they are hiring a lot of people who would understand how to use it.

One day, he received a tablet. And he thought this must be yet another devious idea by his management to make him do more work that doesn't really help him or his customers. A fancy toy, he thought. For the first time in his life, the company positively surprised him. The tablet came with an app that did what he thought the tool should have done all along. As soon as he launched the app it showed him a graphical view of his service calls and parts required for those calls based on the historic analysis of those appliances. It showed him which trucks has what parts and which of his team members are better of visiting what set of customers based on their skill-set and their demonstrated ability in having solved those problems in the past. Tim makes a couple of clicks to analyze that data, drills down into line-item detail in realtime, and accepts recommendations with one click. He assigns the service calls to his team-members and drives his truck to a customer that he assigned to himself. As soon as he is done he pulls out his tablet. He clicks a button to acknowledge the completion of a service call. He is presented with new analysis updated in realtime with available parts in his truck as well as in his teammates' trucks. He clicks around, makes some decisions, cranks up the radio in his truck, and he is off to help the next customer. No more filling out any long meaningless screens. His view of his management has changed for good for the very first time.

As the world is moving towards building mobile-first or mobile-only applications I am proposing to build analytics-first enterprise applications that are mobile-only. Finally, we have access to sophisticated big data products, frameworks, and solutions that can help analyze large volume of data in real time. The large scale hardware — commodity, specialized, or virtualized — are accessible to the developers to do some amazing things. We are at an inflection point. There is no need to discriminate between transactional and analytic workload. Navigating from aggregated results to line-item details should just be one click instead of punching out into a separate system. There are many processes, if re-imagined without any pre-conceived bias, would start with an analysis at the very first click and will guide the user to a more fine-grained data-entry or decision-making screens. If mobile-first is the mindset to get the 20% of the scenarios of your application right that are used 80% of the times, the analytics-first is a design that should thrive to move the 20% of the decision-making workflows used 80% of the time that currently throw the end users into the maze of data entries and beautiful but completely isolated, outdated, and useless reports.

Let's rethink enterprise applications. Today's analytics is an end result of years of neglect to better understand human needs to analyze and decide as opposed to decide and analyze. Analytics should not be a category by itself disconnected from the workflows and processes that the applications have automated for years to make businesses better. Analytics should be an integral part of an application, not embedded, not contextual, but a lead-in.