Custom Development is Dead! Long Live Custom Development!!
Twitter
Facebook
Delicious
Digg
Stumble
Reddit
When you talk about how best to run a software business, we tend to hear one line again and again.
“Build a product! There’s no future in custom development anymore!”
I, myself, have been guilty of recommending this course of action to quite a few people. Lately however, it got me thinking. Why is it that custom development is considered to be a track to failure, while product development is ruling the roost? Aren’t there quite a few product companies which are closing up as well, infact, even more quickly than custom development companies? A product is a risk. If you cannot find a large enough market for it, it doesn’t even justify its costs (ROI). Sometimes, you start out building a product with all the right market knowledge and even interested buyers, but by the time the first beta of the product rolls out, the market seems to have changed and your prospective leads have dried up. So many things can go wrong in the product area as well. So why?
The answer, it seems to me, comes from who you are targeting!
Take a look at the fashion industry. Every person wearing clothes is different, has different body contours, body shapes, preferences, likes and dislikes, as well as budget. Sounds familiar yet? In the fashion space, you have essentially two types of operators.
On are the pre-stitched providers, labels like Littlewood and Zara men (local) and Marks and Spencers, Gap and Old Navy (international) who use a certain fabric and a certain design and then mass produce their products (shirts, tshirts, trousers, jeans whatever) in pre-defined sizes which are just general enough to cover a large percentage of the consumer. Which is why, everytime we go and buy a pre-stitched garment we go through a ritual of trying on every piece to make sure it fits. These labels build their own designs, market them on their own and then even sell them at their own outlets or in partnership with other retailers.
On the other hand, we have the tailors. People who take whatever material you take to them, and stitch them up in whatever design you tell them and at the end of the day, they’ll make sure the clothes fit you precisely. (Barring all the horror stories about tailors which exist).
Now think about it, does the availability of labels and pre-stitched clothing (products) decrease your need to have custom fitting, unique and comfortable clothes (custom-development)? Infact, i’ve heard numerous people claim that if you want good clothes, its better to have them stitched than to go retail, all brands notwithstanding.
Lets dig a bit deeper into it. How to tailors make clothes which are just right for you? There’s a lesson here which would benefit everyone thinking of going into the custom-development model of building software. Have you ever considered why every tailor shop you go to is specialized to a particular segment of the consumer market? E.g. you have tailors for women clothes, for men clothes and for children. Then, you have some who specialize in making eastern clothes for women, while some specializing western clothes for women. The same is for men. You have tailors specializing in western suitings, while others in just shalwar qameezes and a third group which makes nothing but trousers. Then there are shops which aim to provide a one stop shop experience. They have various specialists who help tailor different aspects of an outfit, and then put it all together and deliver as one piece.
The key word, in all of the sentences above, is specialization! Pick a sector, a vertical, a domain and specialize in it! Figure out the needs of the market and fulfill them. That’s where both product based and custom development solutions pay off. One of the biggest problems we see with custom development organizations is the need to fulfill whatever need the customer comes up with. They want a website, you’ll build them a html one, then they want ecommerce support, so you build that in PHP, next they want an inventory system so you build a project in visual basic, then they need automated warehouse management, so you build embedded systems to control their shelves, then they say how about a wireless access the whole thing and you end up doing .NET mobile development… just so you don’t loose the customer. At the end, what is it that you would be known for? Sucking up to the customer? Everybody does that… how are you different?
The objective of tailors is to suit (pun intended) each individual need of the client, while the objective of labels is to suit generic needs of the consumer. The same holds true for software development as well. A product company will never be as good as covering the needs of their customers, just as a custom development company wouldn’t be as good as appealing to the masses.
How you choose which model you are going to follow, depends largely on how you want to satisfy your market. The need is there for both, just go and find it. You will find the market exists and is larger than you ever imagined.
11:08 pm
The reason why software companies in Pakistan are closing up is NOT just because Custom development doesn’t work but there’re OTHER issues which have nothing to do with Custom Development which works great. Here’s my list of those areas of failure:
1- miserable professor/University who focused only on MS based approaches/languages
2- very demotivated/less competent or just plain ignorant Marketing guys who have little idea about internet world. Heck, even CEOs and software engineers don’t get it sooner.
3- no or little satisfactory collaboration and communication between design and development teams. And then did you ever hear HCI related professionals in our companies? Indians even have them.
Custom development works great for many areas and many businesses would/do need it the only thing is my software guy/team/agency doesn’t have clue how to make it work…
Heck, I believe OpenSource software don’t have solutions-for-every-problem either.
Actually I think there’s a good comeback of Custom development on top of already-available solutions (either open source or proprietary) in next few years as more businesses get sick of ready-made designs/developments and there’re signs already. The only need of the hour is that our so-called software programmer starts thinking that just because he got BS and MS degree in Software Engg. doesn’t make him the best person to evaluate, design, market and use the product he’s gonna work on.
As you can see there’re many Facebook applications requests, PHP-developers-required and more importantly the entrepreneurial stunts going on in the country right now. I predicted few years back that many then websites of businesses would be redesigned and it was very true. I personally have worked on websites and applications out of which 90% are re-designs or re-developments. Most of them were to be done from scratch and with bottom-up approach.
11:48 pm
Well your assessment is right but in IMHO only to a limited extent. There is a growing trend back towards custom software but not in the form it was known to be but as SaaS (Software as a Service) model which allows for mass customization and runtime service orchestration. The trend is moving from programming a customization task to ‘sewing’ one by integrating processes and workflows. Such customization offerings are semi-automated to fully automated in the near future (with semantic web services), the challenges will be of hosting a set of services which cater to a target market including the design of modular services, their interoperability issues and ofcourse the hosting itself.
This is not futuristic, ive been developing such services since 2005 and am an end user of it too.
Infact many products are offering features which enables them to be able to be highyl customized via Saas.
2:10 pm
Both approaches have their own advantages and disadvantages, and their own rules of engagement. While off the shelf software software tends to be relatively cheap, because cost is spread over a large client base, developing customized software can be expensive with a longer cash to cash cycle, and it has greater risk because your client base is small. In addition to that off the shelf software can have a recurring revenue stream, while for custom development you have to get into some sort of a service contract with the client to ensure continuous support. However, if you go for specialized custom development satisfying niche market, it can become very rewarding with huge returns. And I agree with Atif’s comments given above. If one is going for a custom software, the best thing to do right now is to go for the SaaS model because it offsets the disadvantages and nightmares that come with custom software development, like deployment and support issues. With the cost of bandwidth going down and the speed going up, that can be the next era of custom development after the demise of disk drives, CDs, and LAN, and maybe Windows. SaaS itself has some limitations right now, but that can change over the few years.
I think in the end what matters is with the type of resources that you have, skills, specialties, and more importantly, the access to market. Both approaches can fail, if you don’t do it right, and both approaches can guarantee you rewarding returns if done correctly.
2:47 pm
ejaz asi: welcome my old friend! its nice to hear your views again after quite sometime. You are quite correct in your views on why software houses fail, however that is the tip of the iceberg only, and something which can be resolved by subsequent trainings.
the model for custom development on top of already available solutions is quite lucrative, and i believe a lot of organizations are already devleoping their products this way. One of the companies i know build only a overlapping presentation layer on top of risk management module built by a foreign company, and they’re doing pretty well so far.
There is quite a large market for website and even web application redesigns and anyone who can do this well is looking at a very rewarding career in the future.
Atif: I may not be as knowledgeable in SaaS as some of the other people here, but i’m wondering how does SaaS becomes custom development? If SaaS, something like even what CDF Snip, is not made for a single client but with a large consumer base in mind, doesn’t it become a product? Consequently, if you develop a software on the SaaS model which will be targeted to only one company, then doesn’t it become a Custom Development effort? In this scenario, you would develop, host and maintain the application till such time the contract lasts on your own premises and the client only gets the funcitonality.
As such, i believe SaaS to be a difference of delivery model rather than a difference of business model? Please corrent me if my thinking is off the mark here.
asim: very well said asim! in the end, what really matters is the access to market, which will pretty well define your model.
12:32 pm
Thank you for your article .It is a very kind gesture of you to part with your wisdom ,am sure this article is going to help a number of people .But most important of all i feel people should figure out what they are interested in and then take their decision on what to do with their career.