Going from idea to MVP is both exciting and bewildering at the same time. A quick search of the interwebs reveals plenty of info about tech stacks, funding, a whole industry of motivational charlatans for startups. But, there’s precious little about the process of finding and working with an external development partner.

It really hit me how absurd the whole process is when a fellow founder mentioned that he’d reached out to 150 agencies to build his startup’s MVP. He settled on one and was satisfied with the result. Nonetheless, the whole ordeal created more questions than it answered. Proposals varied by orders of magnitude.

I don’t have a silver bullet, but here are a few points to help you out as you look for a team to develop your MVP. With a bit of work on your side, you can get much more accurate estimates and save yourself a lot of time and trouble in the long run.

Describe your MVP in concrete terms

Abstract ideas like ‘wouldn’t it be cool to make the Uber of cat toenail trimmers’ are great ways to chat over a few beers, but nobody can give you estimate from that. Such a description is a good place to start for an elevator pitch. Engineers, on the other hand, need to know what you’d like your app to do.

Instead, say that you are looking for an app that lets a client create an account and search a database of cat toenail trimmer—who have their own accounts. Additionally there is an admin account that handles disputes and customer service. Payment will be in-app using Stripe.

That’s enough to give you a ballpark figure that should be within 30% or so of the final estimate.

Think in user stories

When you create personae and user stories, it’s much easier to uncover other features you need.

As the owner of a cat with super long toenails I can find a feline pedicurist based on location, rating, services offered and price.

Showing this to an engineer lets them know that your MVP needs a map integration, reviews and search filters for cat owners. For feline pedicurists there needs to be the ability to edit profiles based on price and services offered.

Know your market

Figure out how your clients are going to access your app. Is mobile a priority or do you predict more desktop users? Is iOS or Android dominant in your market, what about older browsers?

This way an agency can present tech solutions that meet your business needs rather than vice versa.

Racing to the bottom with price

It’s tempting to scrape the bottom of the barrel and go with someone on a freelance site offering to work for $10 an hour. You can win the lottery and get a great engineer, but let’s be realistic.

There are plenty of horror stories with agencies telling clients whatever they want to hear at price that’s too good to be true. These stories don’t have happy endings.

You likely don’t need the most expensive coders on the market for your MVP. Try to balance quality and price. If you cut too many corners at the start, adding new features and scaling is going to be an expensive nightmare.

Culture fit

Be upfront with how you’d like to communicate with the team that’s building your MVP. My experience shows that having a single point of contact for most communication is the easiest. Nonetheless, it’s also important that everyone on the team is visible to the client.

If you’re from a direct and cut to the chase culture, check that the team you hire shares the same values. Language is also a dealbreaker. Be upfront that you need to check the team’s English (or whatever your working language is going to be).

Nobody can do everything

Even the best agency can’t handle every kind of project. If the answer to every question you ask is yes, that should give you pause. If an agency seems to have no core area of expertise and chases after buzzwords like five-year-olds playing football—watch out.

Have a ballpark budget and timeframe

Conventional wisdom dictates that you hide your budget from vendors. In practice, I don’t really see the point of this. If an agency only takes on projects over $50k, but you’re looking to build a $15k MVP it’s better to realize this right away.

The same goes for start dates and timelines. Be upfront about what’s flexible and what’s not.

How to evaluate a proposal

A proposal needs to have a detailed breakdown of how your MVP is going to be built. This is usually illustrated in weeks or sprints for each key feature. Ignore estimates that don’t do this.

Murphy’s never far from any project, and giving perfect cost and time estimates before a single line of code’s been written is tricky. Nonetheless, detailed scoping minimizes risk and gives a solid baseline.

This also makes it much easier to compare proposals. If two agencies have provided nearly identical scoping, comparing prices is much less of a black box.

And the winner is…

If you organize your app idea into user stories and features, you can get more accurate estimate faster. A bit of extra work at this stage helps with business decisions that pay dividends down the road when it’s time to add new features and scale.