Our Process

Over the years, we have watched projects crash and burn. Learning from failure has led us to develop a process we're proud to share. There is a story and a reason behind each step.

These are the steps we find essential for an enjoyable long-term relationship.

Why Understanding Comes First

For the first year after starting PickleLlama, we pretty much exclusively worked as subcontractors for other agencies. Probably the most painful lesson we learned in that time was what happens when no one checks with the end user to make sure that the software being built is actually what they need.

One of our projects came to a very sad and frustrating end when, in the last two weeks of the project, we were presented with a checklist of "must haves" from the client that we had never seen before. We tried our best, but in the end, what we built could only do 20% of what we set out to build.

The happy note is that it still recovered hundreds of hours of time every month and allowed their team to focus on more important things. But it could have been so much more if we had only known what the client actually needed.

This story is a big reason why we don't do subcontracting anymore. Client relationships matter too much, and good outcomes are too important.

That's why every engagement starts with understanding. We produce a detailed report on what we heard you say—so we're on the same page before anything else happens.

Learn more about our first meeting

Why We Try to Convince You Not to Hire Us

There aren't many new ideas. That's the crushing reality of building a business. With 8 billion people on the planet, it's almost guaranteed that someone else already thought of it. That means you probably shouldn't build it yourself.

Now, you might say, "but you are in the software building business..."

Yes. And here's why we still try to convince you to buy something else first.

The Math Usually Favors Buying

Custom software is expensive. Not just to build, but to maintain. A custom solution might cost $50,000-$200,000 to build, plus ongoing maintenance, hosting, security updates, and the inevitable "can you just add this one feature" requests.

Off-the-shelf software is cheap. A SaaS tool might cost $200-$2,000/month. That's $2,400-$24,000/year. Even at the high end, you'd need to run that tool for 8+ years before you'd have spent what a custom build costs—and you get updates, support, and new features included.

We've Watched Companies Waste Money

We've seen a company spend $150,000 on a custom inventory management system when NetSuite would have done 95% of what they needed for $1,500/month. They spent three years building and maintaining something that a product team of 50 people was already improving every month.

We've watched startups burn six months building internal tools that Airtable or Notion could have handled in an afternoon.

We don't want to be part of that story. It's bad for you, and honestly, it's not the kind of work we find fulfilling.

"But Doesn't This Hurt Your Business?"

Actually, no. Here's what happens when we tell someone they don't need custom software:

They trust us. When we eventually do recommend building, they know it's because we genuinely believe it's the right call—not because we're trying to generate billable work.

They come back. Maybe they don't need custom software today. But when they do have a problem that genuinely requires it, guess who they call?

They refer others. "These are the people who told me NOT to hire them" is a surprisingly effective endorsement.

We work on interesting problems. When we filter out the projects that shouldn't be custom, what's left are the genuinely interesting, high-impact projects where custom software is the right answer. That's the work we want to do.

We always recommend buying and integrating first, building second. Before we build anything, we exhaustively research options for solving your problems without hiring us.

Learn more about our research report

Finding the Problems Worth Solving

Most companies have dozens of problems they could throw technology at. The hard part isn't building solutions—it's knowing which problems are worth solving.

We've seen companies spend months building solutions to problems that cost them $10,000/year. That's a bad trade. We've also seen companies ignore problems that were costing them $500,000/year because they didn't realize how much it added up.

Problem identification isn't just listing what's broken. It's quantifying the cost and prioritizing by ROI.

What We Look For

Time drains: Where are your people spending hours on tasks that should take minutes?

Error hotspots: Where do mistakes happen most often? What do they cost to fix?

Bottlenecks: What's slowing everything else down? What could scale if this was fixed?

Hidden costs: What frustrations are driving good people away? What opportunities are you missing?

We don't just make a list. We put numbers on everything—so you can make informed decisions about what's worth fixing and in what order.

Learn more about Problem Identification

When We Build, We Do It Differently

Traditionally, building software has been extremely expensive and time-consuming. Building things in an exploratory fashion to see if they work has been a luxury that even large enterprises frequently couldn't afford.

AI has changed the economics, especially when it comes to prototyping.

The Value of Code: From Caviar to Canned Tuna

Code used to be precious. Every line was expensive to write, expensive to maintain, expensive to change. Companies hoarded code and protected it like treasure.

That's changing. AI makes it possible to generate working prototypes in hours instead of weeks. The value isn't in the code itself anymore—it's in understanding what to build and making sure it actually solves the problem.

The Real Vulnerability of Custom Software

Building production-quality software still takes time and effort—less than even two years ago, but it's not trivial. Production systems still have to be reliable, scalable, secure, and usable.

But the most pressing vulnerability of custom software is actually building the right software.

We've seen projects where the engineering team delivered extremely high-quality software, only to find out in the last weeks of the project that the software solved the wrong problem—albeit very well.

How AI Changes Our Approach

AI doesn't just accelerate building production systems. It allows us to very quickly bootstrap ideas into test articles that prove or disprove our beliefs about the world.

We prototype in days. We test with real users. We validate before we invest in production-quality code. This means fewer expensive mistakes and faster time to value.

Learn more about our implementation process