The fashion business is often criticized for offering unrealistic images about super-thin young beautiful people. Other consumer businesses also get criticized for selling misleading images to people – i.e. how they can become younger, more successful and healthier by buying certain products. But I think there is a business that is even worse than consumer businesses when it comes to offering customers unrealistic expectations: the B2B IT software and solutions business. It’s so bad that one has to ask: is it even possible to buy proper software to run your critical processes, or are companies better off building their own software for their core functions?
The Wall Street Journal published an article by MIT research scientist Joe Peppard saying it is time to get rid of IT departments. The article claims that it’s the IT department that is preventing companies from being innovative, agile, customer-focused and digitally transformed. But is the situation only the IT department’s fault?
IT departments are an obstacle, but maybe with good reason
If you have sold B2B software, you have probably seen how IT departments are often your primary obstacle. You might be able to convince the business management, marketing or HR departments to adopt your software, but then you get to the IT department and the CTO with their policies and skepticism that stop your pitch in its tracks. But perhaps it is not just that those IT people are conservative – they may have good reasons to be skeptical. They have seen too many enterprise software vendors fail to fulfill their promises. Also, to be honest, business people don’t always know what they really need from IT.
The hope of the last ten years has been that SaaS will change the IT business for the better. But now it seems like this hope is probably an illusion. It is certainly easier to buy SaaS services, and the price is much smaller. But often it doesn’t indicate the total costs and the impact on the organization. Currently, many enterprises have very fragmented IT services maps. Many independent services that are not really compatible, and in fact the total price tag goes up rapidly when you adding new services all the time and finding more use for them. And then you need more software components to get those services to talk to each other. Just look how many unicorns have been built just selling data pipelining solutions to get data from one system to another.
A small service that was initially very cheap, or even free with some credits, suddenly has a high monthly fee when you really want to use it. As a friend from a London company told me, “First our quite small company took Salesforce CRM to use for a few hundred dollars a month, but it forced us to use many other Salesforce components and now we pay a million dollars a year, and Salesforce keeps our people happy with fine-dining events.”
SaaS services may sometimes be useful for things like avoiding vendor lock-in issues, but even if in theory you can easily change to another service, in reality it is complex when you have your data in one service and your people have already learned to use it, even if they don’t really like it. Some vendors and system integrators also make open-source promises as a way to avoid vendor lock-in, but if your open-source components are only one small part of the total solution and tied to commercial components, it doesn’t help much.
Could decentralized IT help?
WSJ’s article claims that the current IT department structure doesn’t really enable companies to become digital. I agree with this. I have written earlier that automation and digitization in many organizations is like lipstick on a pig, and real digitization is different from projects where companies try to get modern IT to work with old-fashion processes.
The WSJ article sees it as problematic that current IT departments are like independent business units that sell their services internally, and they are not really motivated by the real business outcomes of the company. This is most probably true as well. The article offers a solution: decentralized IT units in different departments. The writer believes that a more decentralized model would make local IT people in each department be more like partners for the business people.
However, I’m not sure if the problem is really about whether the IT function is centralized or decentralized. I believe the reality is even more complex and the problems more fundamental than that.
Real digital success stories are built on in-house software
If we look at really modern digital companies – for example. Amazon, Google, Uber or Tesla – they are built on digital processes, rather than using software packages to digitize traditional processes. A retail chain doesn’t become a new Amazon by automating and buying new software to run its old-fashioned processes. Real digital processes are the core of modern company, and they are designed to be totally digital.
All of the companies mentioned above have built a lot of their own software to run their core processes. Amazon, Google and Uber are very much running their whole company on their software and genuine digital processes. Meanwhile, Tesla is a good example that a ‘hardware’ company can also be a software company and build its core software in-house. Famously, for example, Elon Musk decided to get rid of SAP and build Tesla’s “WARP Drive” ERP in-house.
This raises an interesting question: should a truly digital company buy software for its core functions, or should it basically build its business on software, and once software is its core, start developing its own software? How many companies have really grown to be on top of their league by using SAP, Oracle, or Salesforce solutions in their core processes vs companies have disrupted an industry by making its own digital processes and software? Those big software vendors often advertise how many big companies use their solutions, but they forget to mention that those companies were big before they started to use that vendor’s software; they are often incumbents of their industries, not companies that grow fast and win market share.
Most probably it is possible to build a successful company on third-party software, but that requires the company to really know what it wants, and have the ability to design its own digital processes, not just try to run their old-fashioned practices with software. Often, this is also easier if they use very targeted third-party software components and API-first solutions for specific needs (as opposed to large packages), and design and manage their processes, data flows and data models in-house.
Who has control: management or vendors?
If a company, its management or business units believe new software packages and some external consultants can solve their business problems to become digital, the project is doomed to fail. Frankly, if you believe that (1) buying a new CRM is the solution to your sales challenges, (2) buying a new automation software makes your processes better, (3) buying a new ERP system make the company more effective, or (4) buying a new HR software makes your employee more productive and happier, you don’t know how to run your business.
You should think what you really need to do to achieve those targets, and then build processes and functions for them, and only then see if it is better to build your own software or use third-party software components to implement them. IT and digital functions must serve management and business needs, not the needs of software vendors. Many business people are under the illusion that buying some software is the easy way to solve problems and get better results.
Many enterprise software solutions fail to fulfill their promises. It is not a surprise that VCs also like the B2B IT business – it has been an easy area to get customers into a trap with investors’ money, and once they are in the trap, it is easy to make them pay. Many software people know the saying that if you make consumer software, you can only sell it when it works. But with B2B software you mainly sell things that don’t work – they are only roadmap promises. But we cannot totally blame the salespeople – it is the buyers who must take back control in IT software and stop believing in marketing hype and sales illusions.