In 1999 I was the 4th person to join a startup called By 2002 we were the largest e-commerce in Latin America. I left the company in 2008 as head of Technology and in 2009 joined eHub, a company that would offer the e-commerce capabilities retailers needed for their digital journey.
At eHub I dreamed of enabling e-commerce in Brazil, It was easy to see the gap between the e-commerce platforms and specialized logistics offered as a service at the time (2009) and the ones "mature players" were operating on. My angle was to create a platform that would give retailers higher success rate on their e-commerce ventures. 


Even though the eHub added value to the players that chose it, it didn't turn out as disrupting as I expected. Later data showed a maturity gap as the main root cause. In a nutshell the majority of retailers didn't have the skill or the proper decision-making-mindset to drive success from the platform. That intrigued me to try to find a way for retailers to leverage business over the Internet without needing to wait until they mature and learn how to do it. To me, the answer were marketplaces. I was confident that the way for e-commerce democratization was a marketplace such as 
Having grown in a large e-commerce and also spent 2 years helping retailers on their e-commerce initiatives I had a unique perspective on why that marketplace model hadn't been done yet. I new the fears and needs of both sides. 
So, on January 2011 I started my company aiming to bridge the gap between large e-retailers and small ones creating an "Enabler of Marketplaces", a startup that would make the marketplace model suited for the Brazilian players. During my pre-ops I tested several assumptions and adapted constantly so that the design would cater to the needs of large e-commerces and small stores and address their fears in a smart and scalable way. At one point I had 2 of the largest e-commerces willing to invest on the company and deploy their marketplaces on that engine.
That would have been "The Market Place Company", if I hadn't met someone...


In July 2011 I met an executive at Itau Unibanco, the largest private bank in Brazil invited me to pursuit a bigger dream. We could change the way people relate to money digitally. So I joined the bank as Internet Superintendent.  At the time, I wrote this post (in Portuguese) stating the we were on the verge of the "Amazon-like" marketplace model in Brazil and urging the large e-retailers to take on where I had stopped and make it happen. I'm glad to see that they did.
I always look for 3 characteristics in a company : Brand, Ability to invest, Autonomy of decision.
Itau has an outstanding brand, huge pockets to invest where it decides to, but as a large corporation, I was certain to find a challenge in the decision making autonomy. Now before I got there I had always assumed that corporate bureaucracy prevents agility... and it turns out that, from the inside, it looked like the other way around ! Lack of agility created the need for more and more controls! So I decided that I would make my "intrapreneurship" to find a way to incubate ".com" agility in this large Bank.
It was hard. At the time any talk about new tech got a "you have to remember this is a Bank" face. Some Agile-Scrum initiatives were at the beginning and getting incremental results in time to market, but that wasn't really agile... the Product Owner was not a proper one... he had no empowerment for decision making... To get agile, I would need to work with different stakeholders and departments, as each one is part of the decision process.


By the first half of 2012 Fabricio Dore and I were discussing an initiative where corporations would try to create "internal startups" and we decided to attempt that at Itau, thus creating what we called "Garages". By the time, Ivan Passos was head of innovation at the Wealth Management division and the three of us started the first Garage, called e-Prev, an attempt to jump up Savings Plans sales over the Internet based on an impressive Design Thinking study Ivan and Fabricio's team had worked on. After 6 months it was clear that our first Garage was not increasing our success metric as expected. So... we failed the first time. 
On a side note, we learned a lot! we captured some agility (1 week after we started we were deploying experiments in production)! and as we were learning people didn't want to "close" the Garage... it took me 2 months to turn it off.
One of the key learnings was that my IT team didn't believe that they could move fast and get stuff done, so I tried another experiment. I set up an unofficial Hackathon in August 2012. Using their personal notebooks and without access to any data from the Bank, we spent 4 hours in a rented loft at Vila Madalena where we created a Facebook App. People didn't know it was unofficial and so thought it was the Banks decision to do it... but didn't even tell my boss until after we had done it... You can check it out here. A month later Harvard Business Review published "The New Corporate Garage" by Scott Anthony, witch reassured us to keep trying.
By late 2012 another chance of trying the Garages model came by. This time was a "expenses sharing app" called Di:Vi:Do ("share" in Portuguese). I had adapted the Garages approach based on the learnings I had on the e-Prev attempt, and we started our second "internal startup". This time I put in place an "urgency incentive": the team would have cash for a 6 months burn rate... no more. They had to show results to get more funding. 6 months later, they had built a great app (first time the Bank experimented with continuous deploy, devops, back-end-as-a-service...), but the app was never marketed, as the business sponsor was so afraid things could go south for him that he compromised the POs decision making ability. So we failed for the second time.


Early 2013, I became technology leader of the "Digital Branch" project, João Filipe Araújo was the business leader and a great partner to work with. I decided to try there what I had learned over the 2 Garages failures... I had a feeling that what was preventing the Garages from being successful were some aspects of the internal culture... So I decided to "hack the culture".
The change was quite impressive. We became one multidisciplinary team with one success metric, the time to market and quality of the project was way better than the usual, some of the best features were not in business requirements, but added by IT people in pursuit of our success metrics. At the go-live date, we still had some items not working and everyone was there trying to the last minute... We had one last shot and we fixed the last item. João called his Executive Director on the phone and got the green light ! Everybody, IT, Business, UX, Operations, everyone was shouting and celebrating that we did it ! That we were live ! such energy, such passion of accomplishment made me feel back in my ".com" days. The Digital Branch broke the sales record on the first day of operations. 
Later that year, I met Rubens Fogli and we partnered to open another Garage a Credit Card e-commerce. I used the learnings from the first two failures and the cultural hackings I had learn over the Digital Branch project on this one. It worked like a charm! We became the #1 on-line Credit Card seller in Brazil!
By the end of 2013 I was called to take on as the tech leader of another project, the "campaign platform". I spent 1 month understanding the project, systems, architecture, and kept asking myself why had they made those choices... I saw the same cultural bias that I hacked before being the root cause of the design they were attempting. So I asked them to trust me in pausing the project for 5 weeks, let me re-assemble the group and deliver 1 campaign in that time frame (it took the Bank 3 months for that sort of campaign at the time). They trusted me and we did "the Garage thing". We created 3 campaigns on those 5 weeks and they are operating over 50 a month nowadays!


On 2014 the Channels Committee, witch I had the privilege of being part of, decided to try this approach on channels. We created a garage for each of our apps and for the Internet bankings. Our apps have gone from 2/2.5 stars to 4/4.5 stars since them, digital (Internet+Mobile) transaction share increasing from 53% in 2013 to 70% in 2016, and our channels market benchmark showed us putting distance over competition!
Three floors in one of our buildings were retrofit to foster creativity, innovation and ".com" culture kudos to Fabricio for the great work there. On those floors technology, business, UX, CRM, marketing and operations were sharing the same workspace for the first time.
Our decision making got faster the moment we moved there! 
By 2015 the ".com contamination" was successful. Our strategy was declared as digital. We were broadcasting a new campaign language and our VP of Retail presented to investors our choice to become "the digital bank".
So we had committed to become digital... stage 1 achieved! Now we had to make it possible...


Since early 2014 I knew that the transformation was not complete. There was only so much we could operate on with our new ".com" culture... We were extremely limited by our systems architecture. I had to solve "the legacy systems issue".
To be able to be digital, channels couldn't remain as "interaction shells" for the legacy systems. Governance at the time required that no data or business rules was to rest in the Channels layer, and it was only by the end of 2015 that Thiago Charnet would finally get the Bank to pursuit a Bi-Modal Architecture. So at early 2014, I had to find midground on this. 
I analyzed previous initiatives related to SOA both at the Bank and in other companies, and I just knew that was not it. Choosing the problem is, in my opinion, half way to solving it. I decided not to try to solve the "reuse of services" and all that inspires SOA. I wanted "freedom" for the front end to handle customer engagement and access the "things" it needed from the legacy. That got me to a ROA (resource oriented architecture) approach, and I decided to frame the problem as "decoupling". 
In June 2014 I was nominated Channels Director, in charge of technology and UX for any customer interaction with our Bank. I invested heavily on trying to decouple the channels, witch required finding not only a technical approach suited for the task, but also experimenting until I got the right concept to design the resources for the APIs. At that stage Fabiola Marchiori and I would spend a lot of time doing and re-doing things until we could find a way to connect and abstract the legacy into new powerful REST APIs resources.  

The decoupling by APIs had worked, but we ran into another more complex issue. Our design needed to be business driven, so APIs were business resources regardless of how they were implemented on legacy systems, and that empowerment comes with a price tag... every time you drive adoption, you'll pay expensive mainframe cycles to serve that increasing adoption. This is going on all over the world, as Rich Bolstridge described in his Akamai post. We needed to go from a Channels layer to a Distribution Layer, and as I said before, by end 2015 Thiago Charnet converged IT strategy to allow this. 
Over the "Bank as APIs" journey I learned that without a disruptive enabling platform, digital would not flourish, as the long time-to-market creates high opportunity cost (as my business partners constantly reminded me) and therefore prevents exploration.
Digital is a 21st century problem and cannot be solved with 20th century tech, and the enabling speed is slowed by how companies struggle in assimilating emerging tech.

round pegs in square holes

It was clear that I had to Pivot. The learnings implied a new framing for the problem to be solved. It wasn't just a technological problem, it was an economical one, and the economical shift was being prevented by the corporations antibodies.
I found the solution for Digital Enabling by designing a "Decoupling Layer that creates economical shift, encapsulating emerging tech from the corporation".
So I started Round Pegs designed to help corporations on this journey.
Pedro Donati.