Blog Post

We’ve reached the end of “build it yourself” software

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

On May 7, 1947, Levitt & Sons announced its first rental homes in Levittown, New York. Of the 2,000 homes, more than half rented within two days. The Levitts were building at a rate of 30 houses per day by July 1948. The company’s assembly chain approach to construction allowed it to meet high demand for affordable homes with a single, replicable solution. The otherwise identical homes could be customized with the interior designs, embellishments and personal touch of the owners.

Today in the tech industry, we’re entering an era of Levittown software, and the change is important for all businesses. Technology companies now build Levittown-style products that are appropriate for a wide range of industries. The base product is quickly deployed, cost-effective and customizable. The companies that “rent” these products on an SaaS (software as a service) basis enjoy the highest standard in software without the costs, risks and tech quandaries of in-house development. This Levittown model is the future of many kinds of software.

“We built it in house” is no longer a bragging point. In most cases, “in-house” is now used to describe backwards software alongside “slow” and “user-unfriendly.” Still, companies resist letting go of archaic proprietary applications. People are used to it and it works OK, so why mess it with it?

This thinking used to work because consumer software and business software used to be similar in that everything was user-unfriendly. However, the rise of mobile devices, web services and web apps created a gap between consumer and business software. The ease of downloading, installing and using consumer apps highlighted just how antiquated business applications had become. Instead of accepting typical in-house business apps as they were, people questioned why they were so much more difficult to use than iOS and Android apps.

So, we’re moving from expensive, custom-made, in-house solutions to cost-efficient, customizable, easy-to-integrate solutions that perform well in a variety of contexts. The build-it-yourself era is giving way to prebuilt software that makes it possible to bring advanced capabilities “in house” almost instantly and get to market quicker.

An obvious cost advantage

Companies like [company]Salesforce[/company], [company]Magento[/company], [company]Yammer[/company], [company]Freshdesk[/company], [company]Asana[/company] and dozens of others are on the rise because they outmatch in-house software on usefulness and cost of ownership. Levittown made modern suburbia possible with all the safety, convenience and community people have come to expect in such neighborhoods. The mass production of quality houses also made home ownership available to millions of additional Americans.

All software is supposed to save time, raise revenue, cut costs and boost productivity, but we now know how difficult it is to meet those goals. New wave tech companies like Facebook, Google and Twitter have to hire thousands of developers with six-figure salaries to offer products that work well. Most companies don’t have the cash or bandwidth to support long development cycles and hire boatloads of developers.

If multi-billion dollar companies are putting all these resources into building products that are their source of revenue, why would non-tech companies think they can build their own CRM, help desk or payment software to that standard?

From build-it-yourself to do-it-yourself

Although build-it-yourself options are less desirable today, the Levittown breed of software has a “do-it-yourself” (DIY) component. You don’t install Salesforce and leave the software as is – instead, you integrate with your systems, tweak the fields, add on additional software and make it your own. All the companies I’ve described have experts who help their clients customize the system like this.

In the world of payment gateways, I’ve found that companies prefer to start with a pre-configured gateway loaded with features that increase conversions. It would be extremely expensive or impossible for companies to build these features on their own. So they use a pre-made solution, customize the appearance, plug in sales tax and ecommerce software, and then sell immediately. The base model takes care of heavy coding the way a Levittown house simplifies building, and then improvements and customization don’t require a software engineering army. Customization is now a DIY process with varying amounts of guidance from the provider.

Some techies might argue that the high performance, scalability and reliability of cloud computing is the key to companies like Salesforce. Without diminishing the importance of those qualities, I think the sweet spot between building it yourself and a prebuilt solution is what makes SaaS products so appealing.

The decision to build an e-commerce platform or global payment gateway is a multi-year, multimillion dollar initiative that can go horribly wrong. The risk versus reward equation points to Levittown software. This trend will allow companies to expand into new markets, offer new services and improve operations faster than ever before. So stop trying to be two businesses. Focus on what you’re good at, and let software companies fill in the gaps.

Ralph Dangelmaier is the CEO of BlueSnap, which aims to be the payments leader in e-commerce.

30 Responses to “We’ve reached the end of “build it yourself” software”

  1. Shed Dweller

    I think the analogy to the builders and those houses is perfect:
    IF you have the standard number of children AND IF you have standard furniture AND IF you work in the correct area AND IF you don’t have any special access requirements, these houses MAY be just barely good enough to keep you protected from the environment for the moment, at least until something breaks.

    I would not buy software with so many IFs and such limited benefits.

    A study I read a while ago covering over 4000 projects with budgets above $7 million found the three biggest causes of failure to any IT project (In-house or outsourced) were:
    1) Outsourcing (Usually due to the outsourcing company not being experts in the problem domain. ie. They knew nothing of the customers industry / Not every business is a bank.)
    2) Non-technical people trying to manage technical people or acting as the project manager.
    3) Big-bang approach. (Otherwise known as a forklift upgrade\Old system out completely unrelated new system in and cut over in 3… 2… 1…)

    We recently had a $200K outsourced project to implement an asset management system to better manage the timing of equipment maintenance outages (a big problem in our industry). We cut the vendor loose 2 years and $4 million later without a working product. They kept trying to change what we wanted to achieve to match the product they had already created while charging us the cost of developing a new system.
    This is a safety critical system covered by legislated requirements so our requirements are not open to negotiation.

    I have first hand experience of some of the biggest names in enterprise software pulling the same crap.

    The improvements we want will now be delivered via a series (approx. 7 at this point) of small incremental improvements to our existing tools all created, maintained and improved in-house.
    No step is so large that the existing tools have to be taken out of service for more than a few minutes at a time and no step will cost more than $20K.
    This incremental approach was not originally used because there was a management philosophy similar to that expressed by the author. A $4 million spanking is kinda hard to ignore, so the philosophy is now shifting.

    To get back to discussing the earlier analogy IF your business processes are not a source of competitive advantage AND IF all you want to do is the same as what everybody else wants to do AND IF you want to do it the same way everybody else does AND IF you are willing to pay a yearly ransom for the software that your business depends on, then fine outsource it. You can partly offset the ransom costs by not hiring anybody to try and improve the efficiency of your corporate tools over time.

    But if you don’t fit any of the above conditions, value the good people in your IT department and let them know they are valued. They, the tools they create and their understanding of your industry and business are some of the big strategic keys to your business’s future success.

    Your business people are just seat warmers if their internal and external customers can’t reach them or they can’t reach the data they need to do their jobs.
    Nobody wants a business to succeed and run as efficiently as possible more than those that are part of it. Nobody has a better idea of how process can improve than those that are part of that process and know where the pain points are.

    Greater efficiency => Less effort required by the workers => less errors due to stress => less re-work => greater productivity => greater profits.

    Nobody is more motivated to help remove those pain points than those that have live\work with those in pain.
    Outsourced companies can leave at any time.

    Outsourced companies only care that the product is so hard to replace, the customer has to keep paying the ransom.

    • Fascinating post – thanks for sharing.

      I particularly agree with your sentiments regarding how much employees of a business are not only fully invested in the success of their company, but also in a unique position to know what their specific pain points and priorities are.

      This realisation is essentially what led me to creating the system that I mentioned above, where the business people who will use an application are the ones who create it.

      There is no longer any need to waste years and millions of dollars on open-ended development projects now that better ways exist. I wish you well with your implementations.

  2. Hugo Chavez

    …this is a sales pitch, right?

    In all honesty, with all the privacy breaches these days, in-house systems(whether developed in-house or purchased/installed in-house) has become a luxury because the company doesn’t scare out every time iCloud, AWS or a big cloud provider gets illegally hacked by criminals or (il)legally hacked by the NSA. SaaS is to start cheap, but if you grow enough get your own stuff and protect your data.

    And yes, I sell software too.

  3. A Commenter

    Good article, traditional businesses and websites absolutely need OTS solutions which is the majority of the market so they can focus on what their products/solutions are.

    Some people repeating comments look very silly below

  4. In my experience, generic software is inherently flawed. The levels of indirection taken to make software generic (and support every aspect of every business) means everything is internally far more complex and convoluted then really necessary.

  5. Robert Sewell

    This article should be posted on The Onion, because I couldn’t stop laughing at the ludicrous claim. There is no off-the-shelf software for every industry or every need in any one industry. You may find quite a lot of OTS that almost fits the need, but it always needs tweaking, and it is rarely tweakable enough.

  6. A Commenter

    Strange, you seem to be living in the past.

    Our company is modernising by replacing all the outdated bloated clunky enterprise apps and support contracts with simple and focussed apps that perfectly meet our needs, most of which are built or customised in-house so we get exactly what we need without any unwanted bloat. This produces massive gains in efficiency and gives us full control over the software we use so it serves our needs in ways that 3rd party software never could.

  7. Alasdair Scott

    It’s not a fair comparison to compare software with house building; we’ve been building houses for thousands of years, but software has been around for less than a century and remains in its infancy despite great advances in that time; you only need to think about how little is understood about how the brain performs high speed pattern matching to see how far we have to go. In this phase it’s important that large numbers of people continue to try out different things at a raw level, otherwise progress will stagnate. I like to compare it with another area of science (medicine) at a similar level and imagine future doctors discussing, for example, the history of imaging and surgery: “Can you believe it? Some doctors in the dark ages actually irradiated people for investigation, and then even treated them by cutting them open with a knife – how primitive and barbaric is that!” Surely there will be similar sentiments by future software practicians about current “established” methodologies…

  8. Open source source software with smart and passionate developers from all around the world is likely to always stay ahead as far as innovation goes, compared to closed, propitiatory software.
    Particularly for infrastructure software, “we are building our own” mantra is almost like a death wish.

  9. Dan Sutton

    Hmmm… I’ve seen companies whose software consists of a mishmash of various things like Salesforce and so on — generally, it’s a nightmare. Most of the online stuff is so generalized as to be useless, and the programmers generally aren’t much good, either: for example, I could drag a better Salesforce than Salesforce out of my arse – and the same goes for all this other crappy online solution nonsense all over the place. Stupid article.

  10. Joe Michels

    I don’t think this model is good for mobile apps that represent a brand. Well made native mobile apps are key to a success of a brand. Two apps I’ve been using are Zillow and SeekingAlpha (with there new iPad app). The Zillow iPad app is superior to their web portal, custom native all the way.

  11. DB Conner

    I’ve been hearing this specious argument for 25 years. Back in the nineties, there was a fad/phase with automatically generated software. ALl software was going to be built that way. There are a handful of companies using generated code as a part of their solution, but the market has determined that “do it yourself” software , whether it is commercial or in-house is what works for most companies and consumers.

    Really a goofy analogy with the cookie-cutter approach. Businesses succeed by differentiating themselves., and their software, which models that business, must do the same.

    Its a joke to characterize all in house apps as user-unfriendly. Broad brush, don’t you think?

    This article is a waste of time, except for amusement-value maybe.

  12. JoePritchard

    Ah yes….let me see…which company out there makes the software that my company uses – specialist design tools for mechanical engineering. Oh…no one. Better keep writing then, hadn’t we?

    The world doesn’t begin and end with ecommerce and such – there’s a whole world out here full of people using software to do all sorts of stuff. I heard the ‘end of in house argument’ in the early 1980s (does anyone remember ‘the Last One’?), the late 1990s and now again.


  13. Jake Weisz

    Amazing this isn’t marked as a “sponsored post” or something, the author isn’t a journalist, he’s a CEO who directly profits from people following the advice of this article. Horribly biased.

  14. Steve Naidamast

    I cannot agree with the comparison of “Levittown” to using certain pre-built applications in place of in-house developed ones. First, let us not forget that Levit himself was a racist as these homes were not available to returning African-American GIs after WWII.

    Second, Levit homes were poorly built and were primarily designed to extract GI funds from returning soldiers with the idea that they could own a home. They could, but again these homes were of rather poor quality.

    It is true though, that many of these homes are still standing on Long Island, NY where they were first built. However, due to the lack of initial space they have been expanded way beyond their original intent. And such as design concepts such as the in-floor heating, which seemed like a radical idea in the 1940s was actually a costly issue when it came to repairs since once something went wrong, entire sections of flooring had to be ripped up to correct a problem.

    I do not believe that IT organizations would want to use such an analogy when reviewing pre-built applications…

  15. so, right. the CEO of a SaaS company is to be believed with this type of announcement? I think not. While SOME software will always be ‘off the shelf’, in house fits the way your company does business.

  16. Gareth Coats

    Loved this article. One thing it fails to mention is that those who take on the role of customizing the ‘pre-built’ solution and then integrating into the clients environment are most successful when they have an in-depth knowledge of the working space and the business domain of the client. That is most true for features that are built in to respond to unique challenges of the clients business domain, their clients expectations and especially for opportunities in the clients domain.

    Knowing your clients broader landscape better than they do helps. Start-ups and others who are dedicated to making work and enterprise life better should bare that in mind.

    • I agree with you there and when I was a consultant/contractor having in-depth industry knowledge was invaluable.

      It is also what makes the systems such as I describe above so successful. After all, who knows a business better than the people who run it every day?

      Up until recently, there was always that “impedance mismatch” between the business people who knew how to run the business and the IT people who knew about technology.

      The traditional solution has been to attempt to bridge the gap, by producing complicated documents to describe requirements and helping developers better understand the business. Both of these activities are highly inefficient and no longer necessary, once you empower the business people with tools that they can use to build their own applications.

      • The “traditional” method has been to use a Systems Analyst, which is often overlooked — usually due to fear of cost, because of ego, or just lack of experience.

        “A systems analyst is an IT professional who specializes in analyzing, designing and implementing information systems. System analysts assess the suitability of information systems in terms of their intended outcomes and liaise with end users, software vendors and programmers in order to achieve these outcomes”

        • Yes, I have worked with and observed numerous Systems Analysts and Business Analysts (similar role, I think) over the decades and many appear to be worse than useless, especially as they are not taken seriously by either the business or the IT people. Given that their whole job is to liaise between these groups, to mitigate the effect of the “impedance mismatch”, they are often underwhelming.

          I suspect the root cause of this is that they are seen as an overhead to a project, rather than an asset and not taken seriously, nor given the authority and access that they require. As you say yourself, the role “is often overlooked.”

          As with any group of people, there are some who are excellent, but far too many seem to be like Tom Smykowski from Office Space: “Well-well look. I already told you: I deal with the god damn customers so the engineers don’t have to. I have people skills; I am good at dealing with people. Can’t you understand that? What the hell is wrong with you people?”

  17. With more and more opportunities for people to build their own business software, as well as the numerous SaaS offerings, it isn’t the end of “Build it yourself”. Rather, it is just a change in the way that things are done.

    I would say that the days of monolithic in-house developments that take years, cost millions and ultimately fail to deliver on their promise are over in many sectors. There is a much better way emerging, which allows business users to design, build and support non-trivial applications themselves. I’ve been working successfully in this space for over ten years and have seen numerous examples of this working well, even though many IT professionals are still in denial and don’t believe that this is a viable way forward.

    In fact, the truth is, it works well and the results are extremely encouraging. I have seen highly complex line of business applications built and supported by business users, with no IT support whatsoever. In one case, I gave a couple of mid-level managers half a day of training on a new system and they went off and created an insurance underwriting, claims management and reporting application. This system consisted of hundreds of screens, extremely complex data-processing and dozens of reports. It was built from scratch, by business people with no prior coding experience (they knew how to do documents in Word and Excel, that’s all). They were insanely productive, building what they needed and rolling it out to their colleagues quicker than they could document their requirements, in a traditional project. It is now in use across their business (300 users) and used to process hundreds of millions of dollars of business every year. They have continued to enhance the system over the years, as new requirements come to light, all without having to rely on their in-house IT developers.

    Other successes include HR systems, retail range-planning and supplier contract management, lawyer time and billing, school management, investment bank trading systems, etc. All built by business people, without a developer in sight.

    I see the future of Enterprise Software as a mix of SaaS and the kind of bespoke applications described above, built and supported by the business users themselves.

    • > They have continued to enhance the system over the years, as new requirements come to light, all without having to rely on their in-house IT developers.

      Except when it breaks and they have no idea how to fix it, or when it needs to interface with custom company-own assets, or when the boss says “Great job, scale it out to all 5000 employees, and ensure it’s available even when the Internet isn’t”.

      Regardless, what you are stating has been going on for decades, end users with little to no programming experience, learn the tools they need to create a tool they need, and often others’ in the same field find that tool useful as well.

      Big, bulky, unruly tools that, when major changes need to be implemented 5 years down the road, you might as well drop it and start over, because no one can figure out why X doesn’t do Y anymore, and their tool “NEEDS X to do Y” or it won’t work!

      “NailsForce wants to charge us WHAT this year? F that, let’s switch platforms…. Oh we’re locked in and have no experience outside of that platform? What do you mean we’ll have to hire developers because the business users don’t have time to rewrite the program? 6-8 months to get the application converted over to using’s platform instead?! We’ll be out of business by then!”

      • Thanks for your feedback. I have to say, so far it has been over ten years since the first customers started using the system I described and none of them have experienced any of the issues you mentioned.

        Worst case scenario: They have to scrap it all and start again, but even so that isn’t a bad lifespan for a new system that lasted 10+ years and cost a tiny fraction of what a traditional development project would cost (around 10%) and was so fast to implement, both initial build and subsequent updates, that they were able to react almost immediately to changing market conditions, rather than waiting months or years for their IT department.

        To me, even that worst case scenario is a massive success, compared to many of the projects I have observed over the past decades. While there have been some successful development projects, many have been terrible failures and it is not helpful to anyone to pretend otherwise, nor is it helpful to assume that “more of the same” will magically improve outcomes.

        As I said above, many people are still in denial about this technology, comfortable in their shared delusion that “it’ll never work”. Sadly for those people, the technology has been live for over a decade and has worked well, much better than I even expected and I created it!

        There is a new version on the way and that’ll really shake things up…

  18. I’ve been hearing these same statements for the last 20 years in software development, and time has proven over and again, they are not true.

    As for ‘Levittown’, the analogy of ‘jack of all trades master of none’ is a better one. If vanilla is all you want, thats fine, buy it off the shelf or subscribe, and get the same as everyone else has. If that fits all you need, then fine.

    But SMART companies, big and small, know developing their own software can give them a MAJOR competitive advantage. They own the IP. They can develop functionality their competitors don’t have. They can implement changes when they want to, and before the competition, they can fix bugs quickly, and not be at the constant mercy of a huge multinational with their own agenda, priorities and seven-figure quotes.

    Also, a lot of these ‘services’ also host data in the cloud, somewhere a lot of clients now reject due to privacy concerns about their data.

    Most in-house developments fail because the management just do not understand software development, hire cheap ‘techies’, or get talked by smooth talking salesmen into choosing the wrong platform, and they turn on it out of their own ignorance.
    This is a pure sales advertisement and should be clearly labelled as such.

    I wont be reading any gigacom articles in future if they publish sales advertisements masquerading as software development articles.

  19. Johnnie Grumblefish

    “All software is supposed to save time, raise revenue, cut costs and boost productivity” ?!?!?! How about “All journalists are should refrain from bombastic, overarching statements and horrible analogies.”

  20. Phil Rack

    Might be somewhat true for Enterprise Software, but not at all true for workgroup or analytical software. Way too much customization required and lest we forget, companies who have and control IP as it pertains to software tend to have an advantage.