The Problem with Buying Vs Building Your Software Services

Have you heard of ELTs, reverse ETLs, serverless databases, data lakes, or MLOps? It’s always been difficult to keep up with the latest best practice or solutions in the IT business. But, in recent years, there has been a profusion of new technologies that might make it tough to stay up with the solutions marketplace. Not to add the new marketing campaigns to coin these multiple words, which might be confusing. This has caused me to consider the best tools for each work. More importantly, how can we as directors, developers, data engineering experts, and data analytics cut through the marketing and hype? 

Let’s be honest: we’ve all been on a sales call when the account executive and sales engineer persuade you that their product is the only tool for you. They will show you prefabricated demonstrations that do not account for the 999 ways the product may fail. This is not to argue that there aren’t a number of wonderful possibilities for new tooling. Whether it’s the current data transformation or the push for low-code services, there are always tools worth deploying amid the sea of services and solutions. We produced this post to provide you with some crucial things to consider while purchasing your next service. 


Before signing any service agreement or contract, you must decide whether to purchase or construct. Again, account executives and sales engineers will not tell you all the ways their technologies may fail. 


If your service is small or new, it may be purchased and then disappear. One issue with tiny start-up businesses in particular is that they may be bought out. This is a significant issue if you decide to utilize one of their services since the corporation that acquires them may decide to do rid of the firm entirely. 

This means that if you depend significantly on this product or service, you may face a costly relocation operation. Migration fees might often exceed $50,000. Hence, although you may believe that a start-up or tiny business may provide a more personalized experience. There are several hazards. 


One of the benefits of low-code solutions and the contemporary data stack is that your team will not have to deal with bug fixes. This sounds terrific at first, and it is, for the most part. If your team encounters an issue, you may submit a bug report, and the service you are paying for will strive to repair it as soon as possible. 

But what happens if they don’t? Everyone thinks that services will patch issues as soon as they are discovered. Yet, in my experience, I’ve dealt with various services where this delayed addressing issues for months. While understandable, teams are overburdened with new features to create and issues to resolve. This may become a big roadblock for an organization, costing them time and money to either work around the issue or just wait for the repair. 

All of this negates many of the advantages that many of these automated new data stack magic businesses tout. This may not be an issue if you have a small development and product engineering staff. You would be faced with the same issue if you had constructed the solution in-house. But, if you have a well-established engineering staff that could have produced a more personalized solution, constructing vs. purchasing may be a superior alternative. Just to give you more power. 


Low-code/drag-and-drop solutions promise the ability to automate processes and save time spent building new features. These tools, in my experience, do speed up development. Particularly if you’ve gotten through the first learning curve. Whether you use Retool or Tableau, all of these tools may help you reduce the time it takes to build new features and processes. 

But they are also less adaptable in terms of what they can accomplish. You can’t go into Retool (a fantastic low-code solution) and reprogram some fundamental functionality since it’s not your code to change. This has always been a key concern when it comes to choosing a software solution. Moreover, the more the software solution attempts to give flexibility, the more time developers will have to spend understanding the solution as well as setting how it works. 

I would argue that many recent low-code solutions are more flexible than their predecessors. I’ve lately tried a few solutions that enable developers to connect API endpoints in the context of a CRUD transaction or a data scrape for an ELT. Even so, there are restrictions, and if you’re a developer or engineer by profession, it may be incredibly irritating when you can’t optimize or enhance service because you can’t access the code. If you are looking to cut down all these problems while buying or building your own software reach out to our Product Engineering Experts directly or visit our website.