Something I’ve seen often is non-technical founders start their project by hiring an affordable (sometimes) agency, explaining their vision and then going along with whatever process the agency follows.

In many cases, a year or more will pass and the project isn’t ready for launch. It’s unfinished and even the core features have bugs.

It’s important to mention that the above process is the wrong way to go about things. Validating with a MMVP first is the way to do it.

But this post isn’t about validation. It’s about how to make technical decisions when you’re already building.

Let’s start with choosing a​n application technology stack. An agency will normally recommend their preferred option.​ But stacks can be like a religion: accepted unquestioned, or chosen for how they make you feel. How can you decide if this is right for your project?

First, ask them for some alternatives and why they chose what they did​ specifically with reference to the requirements of YOUR project.

The reasons you want to hear are:

• time saved due to the stack providing a lot of out-of-the-box functionality required by your project
• stable
• a large component/library ecosystem, which again saves you from spending time doing “undifferentiated heavy lifting” (building things that already exist)
• well-supported and continuing to be maintained

​With answers to the above you’ll have part of the picture.

​You should then do your own validation but trying to analyse the options from a technical point of view is very difficult. Not even the users of these stacks agree!

So we can use social indicators instead: job market, Google searches, and forum posts.

This is not a scientific approach but does give a decent picture of the sentiment towards a given stack.

T​o look at the job market for a stack, put the names into a large job aggregator and count the results.

​For example on Indeed within Australia, we can look at popular frontend and backend stacks.

On the frontend there are:

• 500+ React jobs
• 100+ Vue jobs
• 100+ Angular jobs

and for the backend:

• 18 Laravel jobs
• 25+ next.js jobs
• 25+ Django jobs

Unfortunately, the sample size is too small for the backend options to really help with our decision. You can really only draw from that there are new/ongoing projects that are using these frameworks. Which is good to know.

On the frontend, you can definitely use this info to know that React isn’t going away anytime soon.

​Let’s try Google Trends for your region. I’ll use Australia again so we can compare with the job market.

On the frontend, we can see that React is more popular but Vue has remained steady in its popularity.

Comparing this to the job market we get a similar picture but this data also shows us that the relative popularity of Vue and Angular isn’t changing.

On the backend the data shows a bit of a different picture. Either Django is more popular or it’s harder to use! We can’t really tell from this but again we can see that each option’s popularity is stable over time.

It's important to know that technology stack choice does vary by segment. For example, speaking broadly, .NET is very popular in Enterprise and certain verticals but is extremely unpopular in the smaller agency space.

This is why when performing an exercise such as the above you may need to moderate the results by your vertical (e.g. Finance) to get more meaningful results.

So there are two ways you can perform your own validation and have more confidence around the recommended stack.

Here’s the secret though:

There is no best stack. There are better ones due to good design, good documentation, and a large community, but at the end of the day It 100% depends on the context in which its being used.