HYIPJunction: Project Disaster
Posted on | no commentsThis article is about an outsourced programming project that I’ve been working on for the last 18 months and discusses all the things that went wrong and offers some advice to any would be project managers thinking of outsourcing their programming project.
If you don’t want to read this case study on how the project was a disaster and just want to read my top tips for outsourcing success, click here.
Around January 2006 I came up with an idea to create a super-mega portal out of nothing that focused on the HYIP arena – a underground, murky industry that thrives on the internet, although it’s generally not a very nice place to be. The reason I wanted to create a mega portal in this arena were simple: massive advertising dollars.
I started out as I had done with most projects that I’d done in the past and went about writing down some features that I wanted the site to have. I really pushed the boat out, I wanted it to have everything and be the all singing, all dancing site that would eclipse every other site in the industry (I call it an industry, but the term den of theives might be a more appropriate definition).
Just check out some of the features I was going to create with this site:
- 1. Fully automated advertising system working on an auction basis where the highest bidder wins the advertising position for the week
- 2. Complete trading system that allows business owners to sell shares in their site and let people buy and sell shares. Prices would rise and fall as demand rises and falls
- 3. Awesome rating system with dozens of variables to prevent cheating when people rated the programmes
- 4. News wire system that would let anyone add relevant news items to the site
- 5. Press release system that would syndicate press releases to other sites via RSS
- 6. Articles directory to list articles and let people submit articles to improve the content and the ‘stickyness’ of the site. Users could rate and comment on articles
- 7. Links directory to list relevant links. Users could add links and sponsor links on a monthly basis. The more they paid the higher their listing went. Users could also comment and rate the links
- 8. Every user would get a profile page like on myspace with blog facilities so they could post whatever they wanted
- 9. Integrated chat system to let users come together and chat
- 10. Integrated forum that users were signed up to as soon as they registered on the site
- 11. Membership system that allowed different membership types to earn commission on deposits and upgrades from their referrals
- 12. Portfolio system to allow members to track all their investments
As you can see, the software was going to be the real bees knees; the mutts nuts as it were.
Unfortunately I was not going to have the capital to do this project myself so I took it to my business partner who liked what he saw and we decided to go ahead with the project, taking on some investors (or shareholders) to get the project moving right away.
We spent some time developing the idea further, really thinking about the flow and how everything would work together. Eventually we came up with a definitive 40 page document that we felt detailed every single detail required on the software.
Not only this, but I created a complete mockup of the entire site. Every page we needed was created in static html with every form, every input box and every last bit of text. The entire admin side of the site was created too.
This was the original mockup for the site
So, by April 2006 everything was inplace. The whole programming document was written, complete mockups created and investors were on board to provide some of the capital required. A programming company was contracted and things were looking to go on a roll.
Or so we thought. It was the beginning of the problems.
The site was delivered on 14th July 2006 (just over a year ago…such a long time ago now!) and we went through it and found over 100 bugs. Not only that, but with the site required over 50 changes to the way things worked or things that needed to be added.
By the time we got to mid-August things still hadn’t improved.
This was the first design we had done for the website
A lot of this was down to our doing because we hadn’t explained the features that we required. Some of it was also down to the programmers used because they didn’t understand the hyip arena and obviously had no idea how some other basic scripts like article directories and link directories worked. This was further compounded by them not following the mockup designs exactly as presented.
In a bid to rescue a project that was now 2 months overdue, I flew out to India to work directly with the programmers until things were sorted. However, my plans were somewhat compromised because we delayed in getting some feedback to the programmers and they had already started on another project and it was another 2 months before they could make a start on the bugs list and improvements that we had compiled.
By the time December came around the bugs list was nearing completion and we set about getting a professional design created for the site. The design was completed in record time and with a minimum of revisions – it was quite frankly the only thing that went right with this project. I believe the design was completed so quickly and to my liking because I had worked with the designer on many projects before and he understood my tastes.
Sidenote: Alas, this designer from Romania has since gone AWOL which is a crying shame because he was one of the best guys I worked with and a fantastic designer.
The completed design was sent over to the programmers along with several example pages showing how the design should be integrated. They were under instructions to make the site look exactly how the example designs looked.
I’m not sure if there was any breakdown in communication at this point but when they finally came back to us with the completed bugs list and integrated design the result was a disaster. The design had been converted to XHTML and they had tried to fit it around their HTML framework code…instead of fitting the design into their programming. The site worked well in Firefox but wouldn’t function in IE or Opera. One of the reasons may have been because I didn’t specify which browsers the site should work in.
Fast forward 3 or 4 months and there were still problems with the site. Every time we went back through the site we found more bugs, more changes and discovered more unexpected problems. Not to mention the ongoing issues with the design.
Eventually in April of 2007 – a year since development first began I called a halt to the programming and design integration.
I ended up stripping out all the design code from the files and doing the integration myself. In all it took two solid weeks of work…and you know what? It looked beautiful and worked in every browser flawlessly. To say I was more than a little bit angry that the programmers didn’t do this originally is an understatement to say the least!
This is the final design of the site after I redid what the programmers had done
Now we are in to July and desperate to take the site live…bugs and all. It was due to go live on the 13th July, but more and more bugs were found as we were writing the documentation. It’s incredible that after 12 months of development we are still finding bugs. However, one of the most frustrating things is discovering that something hasn’t been thought through properly and realising that you need some significant changes to a core feature to make it work properly.
Even now, the day we have to launch we are still finding bugs and issues. Even today, we just discovered that there is no way to track how many members a user has referred or no way to deduct affiliate earnings from our income. So many problems, but the site has to be launched. The show must go on, as they say.
====================
- 1. Plan your site down to the smallest of details
- 2. Plan how every function will work, note how it will affect other functions
- 3. Plan out every page that you will need
- 4. When creating your requirements/specification list, include the functions that are required on every page
- 5. Create a full mockup of every page you require and reference it in the specifications document
- 6. Think big, start small. Even though you might want to have the greatest idea in the world, start off getting the core functions done first and develop the rest when that is done
- 7. Keep your development cycles short – think weeks instead of months
- 8. Get regular updates from the programmers
- 9. Provide regular feedback to the programmers – test and test
- 10. When one cycle is complete, take the site live and test and get user feedback
- 11. Test it again
- 12. When one cycle is complete, start on the next development cycle – remember, keep it short!
- 13. When providing feedback, keep it as a bullet pointed checklist instead of long paragraphs. Provide annotated screenshots if applicable
- 14. Accept that there will always be bugs in the software and work on getting them fixed
- 15. If there is a long list of bugs, prioritise them and get 3-5 bugs fixed at a time
These are my top tips for outsourcing success 
If you have any more that you want to add, please let me know. I might make a complete article about outsourcing later on and expand on these tips that I have given.
So…if you were to ask me how I would do HYIPJunction differently…I’d probably say I’d shelve the idea when I had it! LOL. No, seriously, if I was to do it differently, I’ll write about it in another post.
You can check out HYIPJunction for yourself by going to: http://www.hyipjunction.com
Related posts: