Myself and a friend of mine managed to launch a startup (called Hugo) quite while ago but it failed. In this I wanna share my failure experiences. However let's review what kind of startup was it.
Executive Summary
The amount of data in our world has been exploding, and analyzing large data sets -so called Big Data- will become a key basis of competition, underpinning new waves of productivity growth, innovation, and consumer surplus. The increasing volume and detail of information captured by enterprises, the rise of multimedia, social media, and the Internet of Things will fuel exponential growth in data for the foreseeable future.
Big Data, market that will drive IT spending of $220 billion by 2,020, is everywhere and this disruptive phenomenon is destined to help organizations drive innovation by gaining new and faster insight into their customers and businesses and customers.
Hugo offers consultancy and solutions that covers the Big Data Technology Stack:
- Infrastructure: defining the infrastructure that can support the organization growth.
- Data Management: defining the strategy, managing access to extreme growing enterprise information and finding new ways to leverage information sources to drive growth and innovation.
- Data Analytics: getting insights from any source within and outside the organization.
- Decision Support and Automation: predicting future customer behaviors, trends and outcomes.
Hugo access the data from all sort of sources within the company and outside the company (including social networks), makes sense of all the information contained in the data and presented to managers who can access it, analyses it and visualize it. Hugo is envisioned as a SaaS platform to help managers transform, improve and better run their businesses.
Hugo will initially be formed as a consulting company specializing in the implementation of Big Data solutions including actionable analytics, visualization tools, in-memory databases and NoSQL solutions while implementing a SaaS platform to quickly expand its know how to South Asia market.
Mission, Vision and Values
The purpose and mission as a company is:
- To innovate and create cutting edge solutions that solve business issues.
- To create value and make a difference.
Our vision guides us in what we need to accomplish:
- People: Be a great place to work where people are inspired to be the best they can be.
- Customers: together we create more efficient, sustainable and profitable businesses.
- Portfolio: Bring to the market high quality and flexible services and products that adapts to business ecosystem and business needs.
Finally, the values that guide our business are:
- Open Company, No Bullshit: we do embrace transparency; we are not afraid of being honest with ourselves, our staff and our customers.
- Play as team, we want the team feel like they work with Hugo, not for Hugo. We think it's important to have fun with your workmates while working and contributing to the Hugo team.
- Passion, we want passionated people at what they do, people who enjoy what they do and want to make a difference. We need you to love what you do.
- Outstanding value to markets and clients: two things come first, customers and our people; we want Hugo to be recognized in the marketplace for quality in everything we do.
Business Model Canvas
The business model for the first year is based on delivering professional services to gain market awareness, recognition, insight knowledge and ultimately gather as many data as possible in order to build a flexible and robust platform that will allow rapid market expansion and growth in the years to come.
Value Proposition
Hugo helps companies to set up the right and most efficient infrastructure for Big Data, defines Data Management strategy that includes data from all company departments including sales, finance, operations, marketing helping to reshape business design and assemble new information applications that promotes interactivity and increase productivity. Hugo accesses all data that resides in a multitude of sources in the enterprise and over the Internet. Hugo help companies to draw more insight from Big Data Analytics or large and complex datasets while also help to create simulation models that predict future customer behaviors, trends and outcomes.
The following diagram shows Hugo Value Proposition; customers are overloaded with data, information that currently add little or no value to their businesses as it is not presented to right people at the right time, Hugo helps companies to set the infrastructure, data management strategy and the visualization and analytics tools to fasten access to sophisticated insight from any source or combination of data that will ultimately increase productivity, reduce costs and increase revenue.
Marketing Plan
Early in the year, February 2,013, Gartner released Data Warehouse and Business Intelligence Magic Quadrants, two areas that are related to Big Data market:
The study highlights the growth of the market and the acquisitions that leading companies have been doing in order to position themselves in the market. The study also highlights the top five Big Data capabilities are reporting, data mining, data visualization, predictive modeling and "optimization".
The following diagram below shows Hugo's Marketing Plan:
Product and Services
Hugo offers the expertise a high-technology company needs to implement Big Data including data management, visualization tools and in-memory databases. Hugo builds Big Data Applications for a variety of industries. Each application suite is designed based on the baseline data model, predictive models and application adapters that each industry and ultimately customer requires to solve its distinct challenges. All of our applications are based on three pillars that define Big Data Applications:
- Holistic data approach - make sense of all the company related data within the company (all data sources) and external (data stored in partner´s network, social network...).
- Visualization tools - show the right information to right player, converting static data in actionable plan.
- Real Time - making data available at the right time with the current information in order to enable making the right decisions
Lessons Learned
- Lack of Customer Development
- I didn't test the idea, simply I did not get out of the building. The answers to commercialization questions are outside the lab
- Complicated technology which was out of fund of the startup
- The hypotheses was too generic, I wanted to solve so many problems
- There needs to be a separate, parallel path to validate the commercial hypotheses
What else do you think was my mistakes?
I am feeling very excited. At VietnamWorks we are about to launch an amazing feature. We are moving our search into ElasticSearch. Hence we are achieving following items:
- Real Time Job Post Indexing
- Real Time Search
- Increasing the matching hence job seekers can simply search for whatever they want.
- Email Me Jobs now can match better jobs
Besides, it is not just about time ... It is also about search relevancy. To move all of our functionalities at VietnamWorks to ElasticSearch we spent little bit over a month. It was against our Continuous Delivery program but worth the investment. A better search will improve quality of applications as well as creates a better user experience for users. I would like to appreciate all my team members that spent so much effort to make this happen:
- Ninh our Senior Data Architect
- Tan our Database Administrator
- Thang, Son and Lan our Super Senior Developers
- Eduardo our Head of Product who saw the value of it for users and trusted us
Moving to ElasticSearch for a big firm like us was a big and risky move. We did it to ensure our users, experience a better search and job posting. Enjoy the better search.
The Shawshank Redemption is one of my favorite movies. Thanks to Andy Ng for introducing it to me few years back. It has a scene on the rooftop which represents clearly a true definition of a leader.
The leader of the scene, he puts his life in danger just to provide beer for his colleagues (not cell mates), the leader didn't even drink, just sat down and enjoyed seeing his colleagues are relaxing like a free men.
I got inspired by this scene. Especially I believed there is a serious need to change the leadership style of IT knowledge workers. I had a team back then, I started to practice bunch of things to ensure I can lead my team like the hero of Shawshank. It was quite good then I published an ebook which is like 40 pages.
Then I moved to VietnamWorks, where I continued to lead a team of very smart knowledge workers. My colleagues are experts in these areas:
- Product Management
- Content Specialist
- IT evangelist
- Data Architects and Data Administrators
- Cloud Architect
- Very Senior Developers
- User Experience gurus
- CRM and ERP experts
- and IT enablers
I used the same patterns and I improved them. Now I revised the booklet and you can find it in here.
Please let me know what do you think.
There is a war happening. The art of this war is to retain your talents (or so called rockstars). It is pretty difficult to find best of the bests. Hence, when you land great employees, especially senior hires, you have to do whatever it takes to retain them. How?
- Give them the best compensation as best as you can ,
- As soon as you could afford pay them higher than market,
- Continuously talk to them, you need to have a good relationship with them,
- Try to get feedback anonymously like a survey every quarter.
Above items will help you to retain them but what are the signs that are indicating you are losing your rockstar:
- They've no further complaints. It means they believe they should put their energy somewhere else,
- Your rockstar used to be very active in share his or her achievements in emails also sharing what he or she has read. Suddenly there is no more email, specially if it is after 6 months joining,
- When you try to offer them a raise, it is not making them excited,
- Rockstar is avoiding to have a chat with direct managers
Keeping rockstars is an art and each company needs rockstars for their glorious success.
While ago I posted a message on my linked which had above message: "work hard in silence, let success make the noise!". I got over 28 likes. It seems people liked this approach despite I am considering myself as loud mouth.
It made me think about what is all PR and marketing activities then? Imagine we have a world with no ad or public campaigns, is it gonna be a better world? I am reluctant now.
After decade of working in IT on various perspectives mainly technology, development and product; I have been extremely interested in new combination of industries.
I love technology which is my strength. These days I am working closely on different sort of technology. But I have learned something very important that technology if it is not solving any customer problem, it is absolutely not valuable. This is the reason I started to learn about product and how manage a product and I wrote a post which went quite viral on social networks.
But product is not there to solve the customer's problem, then who does the customer development? How to increase sales? Then I find out about new Sales Model. It introduced me to a new world which is called marketing.
I started to read all the books which are recommended by Steve Blank. I read each of them more than once. Crossing the Chasm made me understand that there are repeatable patterns in early stage companies.
Let me share a true story that I benefited from all of above skills:
At VietnamWorks there is a service called "email me jobs" which is basically a job alert, it means you subscribe for your desired job, as soon as it is posted you will get an email of jobs which has matched with your criteria. When I joined in October of 2013, "email me jobs" was driving solely 0.74% of traffic. By combining various skills in technology, product and marketing now "email me jobs" is bringing over 26% traffic which has increased the number of applications in each day over 4500.
Hence, I decided to position myself as a technical guy who loves technology, product and marketing who you can count on:
- generate more revenue,
- increase customer satisfaction,
- increase efficiency in business processes
- build a champion team including content writers, ERP experts, CRM gurus, a web product expert, amazing developers, network support team, User Experience members and data team.
Goal Management offers:
- Create a goal where each goal is assigned to an individual sales person or a team.
- Tracking important business data, such as revenues, customer services cases, or phone calls
- Measure the results against the targets or stretch targets
Thanks to Quyen now we can even do more sophisticated things such as custom sales force automation, real time alerting and predictive analytics.
In my previous post (How to test your startup idea?) I talked about a cheap way of finding out is your business gonna work out or not.
In this post, I wanna talk about the next step which is how to move to next level of online marketing.
Unbounce has launched an interesting guideline which is called Noob Guide to Online Marketing.
This guideline has covered 8 areas of marketing and also comes with a Gantt chart style of activities as well.
These areas are:
- Social Media Marketing (SSM) is the new darling of the marketing community. In Noob guide, there are few guidelines on some social networks such as Twitter, Facebook and LinkedIn as well as few strategies such as timing the tweets, seeding some fans, networking on LinkedIn and etc.
- Email Marketing at Noob is not just about sending emails, it will help you to do A/B Testing your emails, setting up a drip campaigns for acquisition, education and retention.
- Lead Generation is required before doing any email marketing. Who do you want to send email if you do not have any email? You need to give away something if you need some emails.
- Organic Search Marketing is one of your most important elements. SEOMozis the best community on such realm.
- Conversion Rate Optimization (CRO) is the step making you save your money. Imagine you brought in traffic but it is not enough. You need to convert them.There are various approaches such as 5 second test.
- Analytics is needed to achieve above CRO. You need some custom dashboards in Google Analytics as well as tagging some important events.
- Content Marketing is the king. Write an ebook, seed your contents on social hubs and more strategies.
- Pay Per Click (PPC) is the #1fasted way to get traffic IF you do it right and do it well.
Our content specialist (Truc) who is a very creative person, created a presentation about #WorldCup.
This presentation became one of the best ones and it has been share in ourSlideShare account. This is the direct link to the presentation. After 10 hours something amazing happened. SlideShare posted our presentation on their homepage and this is their tweet:
Now 2 days before posting it. On Tuesday, 10th Jun 2014,
I had a chat with Truc in a coffee shop at working hours. She complained to me that we are quite slow in creating contents there are so much barriers (made me to think about Agile Marketing). I was feeling a bit lost, what to do or what to say.
I decided to do something crazy. I suggested to Truc why don't we use the opportunity of #WorldCup and create some content about it which should be related toVietnamWorks. The crazy part, let's do it without any review, and yes. It happened. This post went to the homepage of SlideShare.
As a conclusion, do not limit your creative content specialists. Do you agree?
It has been quite sometime we are looking into ElasticSearch. Data team of VietnamWorks is doing various activities which is being led by Ninh our Senior Data Architect.
We are moving our whole search system into ElasticSearch hence before making such decision we decided to have a benchmark, ElasticSearch vs MySQL. Ninh setup a load test environment and scenario in our lab at VietnamWorks.
MySQL server: 4 processors ( Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz, cache size 15M) 32G RAM,
ElasticSearch had 1 master node: 4 processors ( Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz, cache size 15M) 8G RAM and 1 data node: 4 processors ( Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz, cache size 15M) 16G RAM
Test Dataset was over 429,700 jobs, on MySQL has stored in 4 tables with 1,822,000 total rows and 1.8G total size and ElasticSearch stored in over 1,285,500 documents (has nested documents), with size 2G.
Our query is quite complicated and it is 3kb file. This query is a search for online jobs with conditions on 14 fields, sort on 4 fields (2 non-stored fields).
Ninh used jMeter 2.11 with following configuration:
- Number of threads (users): 100
- Ramp-up period (in seconds): 1
- Loop count: 2
Run each request separately instead of running both MySQL and ES request in same thread group.
Ninh concluded that ElasticSearch is quite faster than MySQL in our complicated massive query.
- Average response time of ElasticSearch (the best) is faster 33% than this of MySQL.
- ElasticSearch is more stable than MySQL 48% (Std. dev.).
- ElasticSearch has no error while MySQL has 1%.
- ElasticSearch has 33% higher throughput than MySQL .
Thanks to Ninh, we are confidently moving to ElasticSearch. Will keep you all posted ;)
While ago we had our innovation time off and we were thinking about usability of our web site and landing pages of email me jobs. Then our User Experience mentor, , introduced a concept to us which is called: 5 second test. Five second tests help you understand people’s first impressions of your designs. We used this test and it helped us a lot to improve our usabilities.
Why run five second tests?
- Test our message
- Test first impression
- Collect users feedback
As our User Experience mentor, Lanh, always says, build a product that users fall in love with it. 5 second test helped us a lot at VietnamWorks.
At VietnamWorks we have over 500 users located in Saigon and Hanoi, Vietnam. We have 406 workstations, 12 servers (in our internal server farms), 7 printers, and a lot of VOIP devices, smart phones and etc.
- There are a lot of concerns for such massive infrastructure.
- There are a lot of requests and supports tickets
- What is status of each device on security
- Monitoring our printers ink
- We care about Green IT. We need to monitor our much power we are using and how to save it.
- and you can imagine a lot more.
Our IT Evangelist, Ho Vo Minh and his team members (Tan, Minh and Quang), introduced an amazing platform which is called SpiceWorks. It is an amazing software. Now we can address all those issues.
We even can create our custom reports for ticketing systems, which generates a report like this:
Last week at VietnamWorks.com we were brainstorming about a logo for our amazing feature called "email me jobs". I was as usual very picky and sensitive about its design.
I was trying to persuade my team members that logo matters a lot and it can even help to make the feature or product more successful. In this brainstorming session, we read various articles, I found this one pretty interesting and I highly recommend it. I would like to have a short summary of why logo matters:
- Trust - A good well thought logo gives the subject a professional aesthetic which it increases the trust between parties.
- Remembering - A good logo makes people to remember your logo, "a picture worth a thousand words".
- Referral - A good logo also lends itself to word-of-mouth referrals and your perceived reputation.
- Offer - Logo can simply represent what your service or product is offering.
So for our "email me jobs" we drafted following logo:
These are the reasons we came to this logo (Thanks to Ly for his nice design):
- The gun zoom mode represents vantage point of job seeker which is targeting an email. email me jobs is a service that helps job seekers to create their own tailored content in an email format,
- There are 5 colors in this logo which each one has its own meaning and reasons which are:
- Green color is the first step, which mean subscription to "email me jobs";
- Dark blue color is our responsibility to find your matching jobs and email it to you;
- Gray color is the time job seeker gets the email, clicks on his desired job.
- Light blue color which is our brand color, is the vision and mission of VietnamWorks. This step means apply for your dream job, VietnamWorks is committed to find your dream job.
- Orange is the email itself, representing the color of dream and hope.
What do you think? Does your logo matter?
When I was just getting started as a CTO, I had a stupid way of thinking about employees. I thought that I was pretty good at doing a large number of things and I could do most of my employees' jobs better than they could. Of course, I didn't have time to do everything, so I would let people do their jobs even though I thought the results would probably be better if I did them. I would look at a programmer and think, "I could write that system faster, but I don't have the time." I would look at a sales guy and think, "I'm better at selling our product than he is, but I don't have time to go on every sales call." I'd look at the receptionist and think about my excellent phone manners.
And I was probably right. After all, why would anyone really great want to work for that kind of boss? When I became head of IT at VietnamWorks.com and also in charge of running our products (vietnamworks.com, ERP, CRM and Analytics systems); I made a new rule:
Everyone who reports to me has to be much better at doing his or her job than I could ever be.
I am pretty good in leading and being creative product developments. I even wrote an article "How to get a realistic product roadmap" which became quite viral. However I hired Eduardo Mora. Eduardo, Head of Products at VietnamWorks, is great at planning, and running, a dozen complex product teams at the same time. After a while, I strongly believe in what I said.
Everyone who reports to me has to be much better at doing his or her job than I could ever be.