Having spent years on both sides of the table shopping for web development services/clients, I’ve both seen and contributed to the dramatic price swings you see when requesting proposals for web contract work. I find it incredibly ironic that internet services make most marketplaces much more transparent, but not the market for internet services. I want to make a few remarks about why this phenomenon exists (nothing comprehensive, but hopefully illustrative), give you a few bottom-line price ranges to justify my search-engine-optimized headline, then introduce an idea that I think might help equilibrate expectations and remove some of the mystery surrounding the costs of building your web application.

Why the huge ranges in cost estimates?

You’ve probably read about (or experienced) quotes for web development that ranged from $10K to $100K+ for the same product with the same requirement specifications. And there are myriad stories across the internet talking about building websites for $1000 or less. What gives?

1. Web Sites and Web Applications Are Different

Unfortunately there isn’t a great black-and-white distinction between these two types of internet “products” but there is a huge cost difference between building a website like joetheplumber.com and dynamic web applications like Facebook or Twitter. One heuristic you can use to distinguish is that websites primarily display static content that a website administrator or designer has generated, whereas web applications are typically processing and rendering content that users have generated. And web applications typically allow users to do more interesting things than just search and view content. As an example, you’ve probably used Wordpress (a web application) to compose content for your website.

Now, the cost of a website is driven by the design and the content, because you can easily find out-of-the-box solutions to handle the content delivery (servers/hosting). The cost of a web application is driven by the complexity and scope, because you will likely need customized server and view code to process/render the user-supplied content or actions. Writing this code requires a highly specialized (and usually expensive) skill set.

You might have website pricing in mind when writing your RFP, but requirements that fall into web application territory for implementation. My services and experiences are limited primarily to web applications, so the rest of this post will only be applicable to custom web application solutions.

2. The Internet Looks Easier Than It Is

You’ve probably never encountered a web application that cost less than $50-100K to build. Why? Because early web applications (MVPs or low-budget projects) are limited to a small audience for testing purposes, and only a small minority achieve widespread adoption. Before web applications scale they always go through multiple costly iterations, and at a certain point require full-time developers that will clock in around $87K as an average annual salary depending on your location.

When you hear people talk about building an application on the low end of the price spectrum your selection bias is going to make you think of Twitter on the cheap, because you’ve probably never encountered any of the thousands of low-budget projects that never got traction and warranted additional investment. High-end proposals typically build in additional time for iterations, additional resources for design/usability, and additional scalability measures that low-end proposals will ignore.

3. Long-term Versus Short-term Thinking

In any product as complex as a custom web application, there are literally thousands of corners to cut. All you will see as the product owner is the front end, which frequently represents the least expensive part of the web application. This means that the cost drivers are all happening in a black box. Low-end proposals are going to avoid things like robust testing frameworks and best-practice code design principals because the client won’t see, understand, verify or value them.

More established freelancers and development shops insist on building these into their pricing structures, because in the long term both maintenance and “reputational” costs depend on these hidden but critical components.

4. Complex Products Are Difficult to Estimate

Unfortunately it is hard for humans to estimate time requirements for complex tasks even if you’ve done it over and over before (see Planning fallacy). Development companies that chronically underestimate will eventually go out of business because they have inflexible cost structures. What remains are companies who chronically overestimate (hence your high bids), freelancers who can more easily survive underestimation (by losing sleep/time/opportunity but not money), and offshore companies who have enhanced resilience due to comparatively low labor costs.

So Bottom Line, How Much Will It Cost?

Of course, it depends. But I’m more prone than others to toss around numbers as a way of setting rough expectations, and this is generally what I see in the market right now. For a straight forward MVP with the average scope/number of features (e.g. a simple social network), you’ll probably pay $10-60K for a good freelancer or small team, and $50K-150K for a brick and mortar web shop based in the U.S. You’ll probably pay 30-50% more for a mobile app, because it is more specialized.

That being said, results will vary and there is still too much variability/opacity for my taste. I’m interested in building a simple polling tool that lets people submit anonymous project costs with simple complexity/satisfaction selections. If there is any interest in something like this pipe up in the comments and I’ll put something together.