Everybody has a stack. Almost nobody has a list of jobs.

The advice is everywhere and it always sounds responsible. Build your tech stack. Get your systems in place before you scale. Find the right platform and the growth follows. Owners repeat it to each other like it's settled, and the whole conversation stays parked on which product to pick.

Go looking for what people say a year after they picked, and the tone changes fast. One solo founder wrote his up this spring: fourteen paid subscriptions before a single paying customer. Analytics, email, hosting, error tracking, SEO. His own description was that he'd been "building infrastructure for a business that didn't exist yet." When he finally ran the list against one question, six tools survived and eight didn't, and three of the eight he never replaced with anything at all. Those three had only ever made the work feel underway.

I see the same shape from the other side of the desk. Owners arrive with a genuinely impressive pile of accounts, and somewhere in that pile is a finished thing that has never earned a dollar. A book that's written and not being sold. A course recorded and sitting on a drive. A quiz built and never launched. The asset exists. What it was supposed to be responsible for was never written down anywhere, so nobody ever noticed it wasn't happening.

And those accounts rarely sit anywhere near each other. The domain lives with one company, the email with another, the payment link somewhere else again, the content on a drive nobody has opened in months. Every piece was bought for a good reason on the day it was bought. Put together they don't add up to a path a customer can walk down, which is exactly why the finished thing stays finished and unsold.

Tools matter. The order is where this advice falls apart. It treats the purchase as the decision and the setup as a formality, when the call that actually determines whether money moves is the one almost nobody makes out loud: what is this thing responsible for, and how will I know in thirty days whether it did it?

I know that order intimately, because I had it backwards for years and finding out cost me over twenty thousand dollars.

The two questions, and the page they live on

Back when this was a lunch-hour business, my buying rule was closer to a mood than a rule. Something felt missing, a program or a platform promised to fill it, and the purchase itself felt like progress. I've written up where that habit led and what it cost me. What I want to hand you here is the thing that replaced it, because the replacement is boring and it works.

Nothing gets bought in my business now unless it can answer two questions. What does this replace? What does this produce?

Replace means something leaves. A subscription I stop paying, a step I stop doing by hand, a tab I stop opening every morning. If nothing leaves, the purchase is an addition, and additions need a much higher bar than most of us give them. Half the bloated stacks I look at are three products doing one job because nobody asked this question on the way in.

Produce means a specific thing shows up that I can point at. A booked call on the calendar. An invoice that goes out without me remembering to send it. An onboarding sequence a new client actually receives. "Saves time" is not a product. "Helps with marketing" is not a product. If I can't finish the sentence "this puts a blank in front of me every week," it hasn't answered.

Both answers live on one page, and that page has four columns.

• What it is
• What it replaced
• What it produces, in one observable sentence
• The date I look at it again

Everything I pay for has a row, including the things I love. New purchases get a row before the card comes out. If I can't fill columns two and three straight off the sales page, I don't buy it that day, and most of the time the want passes quietly on its own.

The review is scheduled, the same way my prices are. It sits on the calendar as a recurring appointment, so the stack gets examined on a date I chose instead of on a day I happen to feel frustrated with it. At each review every row has to still answer column three. A row that can't is either given a real job that week or it comes out of the budget.

One more rule, and it saves the most money of any of them. I don't buy a tool to solve a problem I haven't done by hand yet. If I've never run the sequence manually, I don't know what the steps are, which means I can't tell a platform what to do and I can't tell whether it did it. Doing it by hand is slower for about two weeks, and it's the only way the automation ends up pointed at something real. It's the same reason the CRM is the wrong place to start.

All of it rolls up to the one number I check on a normal week: cash collected against active clients. A stack that's working shows up there. A stack that's decorative doesn't, however good the dashboard looks.

When this test will talk you out of the right thing

Run these two questions too early and they will clear out things you actually need.

If you're genuinely new and you don't yet have a repeatable way you deliver, then almost nothing you own produces anything, because there's no machine yet for it to be part of. Ask "what does this produce" in month two and the answer is "nothing yet" straight down the list, and you'll strip yourself back to a laptop and talk yourself out of the one platform that would let you take a payment. In that season, ask what would embarrass you in front of a paying client. Buy that. Do the rest by hand until the pattern shows up.

The second exception is boring infrastructure. Your domain, your bookkeeping, your contracts, your email deliverability. Ask what those produce and you will get nothing back on any of them. They earn their place on a different question: what breaks if this disappears. Anything clearing that bar stays on the page with a note beside it, and it doesn't get dragged into the quarterly review to justify itself all over again.

And if a tool truly replaces something and produces something, but the plan you're on can't hold the volume you've got, the honest fix is a bigger plan. Running real client work on a tier that drops things is an expensive choice wearing a cheap costume.

Your stack is a payroll

Every tool you pay for is on staff. Give each one a job you can say out loud, then check on a schedule that it's still doing it. The ones with no job will name themselves.

Karena Calhoun

Founder · WarKry Consulting Solutions

I'm Karena, an introvert who loves to build things, devour good books, eat chocolate and seafood (just not together, please), and help people build businesses that actually work without working them into the ground.

I started WarKry Consulting Solutions in January 2018 while working full time in high-level management, running four departments. I was building on two 15-minute breaks and a lunch hour. Newly married. A new blended family. Deeply embedded in my local ministry. In March, two months in, I lost my dad.

The next four years were not a straight line. My job and my family had to come first, so the business went quiet for stretches at a time. I would start, then stop, then start again. What never went quiet was knowing I was called to build this. Over those same four years I spent more than $20,000 on coaches, consultants, and "business in a box" programs, and most of it did not work.

What I got back was hard-won clarity. I learned how to navigate, how to pivot, and how to build something real. That is what WarKry does now. We provide strategic solutions that impact.

Visit warkry.com