Should we work with a software agency or a consultant, and what exactly is the difference?

If there is no one on your internal team with a technical background, do not negotiate directly with an agency. Keep your strategy clear: 1) define scope with a consultant first, 2) request proposals from at least three software firms using those specifications, and 3) have the consultant review and approve the delivery criteria and penalty clauses in the contract.

Last year we went straight to a software agency with a 400.000 TL budget for our logistics operation. We wrote the specs ourselves using layman's terms. They said it'd take 5 months, ended up taking 14, and scope creep pushed the final bill to 650.000 TL. The consultant fee felt high initially, but we learned the hard way.

Don't blindly trust every consultant out there either. Some consultants run secret kickback deals with certain software firms. They make it look like they're preparing an objective report, but behind the scenes they frame their preferred agency as your only viable option. Vet their references very thoroughly.

A software development agency's main job is setting up the architecture and writing the code, but most agencies only do a superficial business analysis based on what you tell them. A consultant, on the other hand, defines the technical acceptance criteria—from normalizing the database schema to ensuring the API architecture is flexible. When the code is delivered, they also verify the documentation and make sure you retain ownership of the source code.

we faced a similar dilemma and when you can't clearly define what you want, the software firm understandably just builds what they think you mean and then both sides end up frustrated. if you have the budget, a consultant is definitely a lifesaver and will cut your headaches in half.

even if u hire a consultant make sure to pick someone internal to oversee everything but otherwise the consultant just writes reports and the agency writes code but the employees won't use the system, so the money goes down the drain anyway.

Have the software companies you're talking to previously completed a similar ERP or inventory management project in the packaging or manufacturing sector like yours? If they have industry experience, it might work out without a consultant but if they've never worked in this sector before, the process will be very painful.

To sum it up: a consultant translates what you want into developer language and protects your interests, while the software agency codes the project and makes it run. If your budget allows for it, having both is the safest bet.

Correction: I misremembered the figure, it was a bit lower.

the discussion got scattered let me summarize. btw in legacy systems, the most expesnive thing is unknown dependencies.

You're right, I've been down that road too. The biggest time-waster for us was not knowing who had the final say.

Code without setup documentation isn't yours, even if you have it. Just leaving this note, it might be useful.

I went through the same thing.

Let me share what happened to me; it might be useful. Payment schedules should be tied to project phases, not calendar dates.

Just leaving this note it might be useful.

Correct. When making a decision, first look at what data you have on hand.

Good luck with that.

Absolutely. If I were to add anything: Hasty decisions become decisions you have to fix six months later.

Thanks, that was the answer I was looking for.

The discussion got scattered, let me summarize. Start with a small trial; don't commit to everything at once.

I'll try it.

The cheap-looking path usually ends up costing more later. Payment schedules should be tied to project phases not calendar dates.

The system lives on after delivery; maintenance is a separate line item. Good luck with that.