A few times in my career, I’ve implemented new software systems.

Every time it’s hard.

Every time you do, you risk making a bad decision or breaking something along the way.

In almost every business (I’d say all businesses, but I’m worried the ACKTUALLY people would get me) there is what you’d call knowledge debt.

These are the gaps that happen between recognized knowledge versus unrecognized knowledge. Unrecognized knowledge is the things you know, but don’t realize you know, thus you don’t plan for them as you shift to a new system.

Knowledge debt can appear in a few ways:

1. Gaps between the new system and the old
2. Lack of documentation of current processes
3. Failure to grasp the new system and its intricacies
4. Failure to identify all integrations between systems
5. Failure to understand user needs

When implementing new software, this stuff can be downright dangerous.

If you have a tight timeline, it’s so easy for that timeline to be thrown off and put the project into a tailspin. This is especially the case when you have many parties involved with competing timelines.

So, to avoid these gaps (as much as possible), I like to evaluate two things:

1. Performance
2. Features

Performance

Performance criteria are the high-level business operations things we look at. We’re not looking at the cool things the software does, but instead we’re looking to make sure it meets our minimum criteria.

If the software isn’t usable, scalable, priced right, easy to implement, or integrate, it can be pretty much useless.

Software companies like to sell on features, but without the right performance, the features don’t matter.

For each of these criteria, you can rank the tools on a scale, for instance from 1-5, with 5 being the best. The rankings can then be added up to give a total score for each tool.

This can help you see at a glance which tool might be the best fit for their needs.

However, you might want to weight some of the performance criteria higher. You can adjust your scoring scale so that each criteria has different “best” numbers.

It might be that you value automation, care a little about scalability, and don’t care at all about implementation. So, you could make automation on a scale 1-8, scalability 1-5, and implementation 1-3.

The great thing is it’s your preference, so there is no wrong way to do it other than not doing it at all.

Now let’s talk about the performance criteria.

Usability

This being first is no mistake. Software’s usability is a big part of how well it will be adopted. A few questions to consider:

1. How intuitive is the interface?
2. How customizable is it to specific company needs?
3. What are you able to do on mobile? Will this work for your users?
4. Can users reach out to support to answer their own questions?

Integrations

If you’re using ERPs, accounting software, or other business tools, integration with those tools is key.

If integration is not possible, you need to know the new workflow BEFORE you make the switch. I’ve heard of companies making the switch without knowing this and then realizing getting the data into their main system was too hard and having to switch back.

I’d put a high priority on systems that integrated over those that don’t.

Scalability

How fast you’re growing determines how important this is. There are a few things to think about:

1. Can they scale with your volume of transactions? How will that impact cost?
2. If new processes are needed with growth, is the solution able to support those?

Pricing

Cost is always going to be a consideration. I like to hold off on this until the end, if possible.

Then, once we understand the clear ranking of solutions, ask if it’s worth paying for the “nice to haves.”

Also, as mentioned with scalability, understand how the price changes as the business does.

Something often forgotten is that a high baseline could be bad news if your business shrinks. You want pricing that can shrink if something goes wrong.

Implementation and customer support

Every new software implementation that was hard resulted in it not getting fully implemented.

Before you sign on any dotted line, you need to understand how you’re going to implement it and who’s going to be a part of the process.

Training and support from the software provider is nice, but if people on your team aren’t committed to the process it doesn’t matter.

Also, consider customer support after the fact. How much access do you have and how quickly will they respond? You can learn this from looking at customer reviews (which we discuss next).

Customer reviews

I know I made this a criteria, but I don’t actually score this one. I use this to inform:

1. How I rate features
2. What questions to ask

It informs how I rate features because the demo will make the features seem bulletproof. But, if real users are saying it’s hard, it’s likely hard. Just make sure you look for themes and do not base this off individual reviews.

By understanding current customers’ experiences, you can get insights into the performance of the software in ways that aren’t immediately obvious. This helps you frame questions for your context so you don’t miss anything.

Demos are good at leaving out key weaknesses and we want our questions to force them to address them so you can make a more informed decision.

Features

Once you’ve passed the performance criteria, it’s time to look at the features.

One of the biggest mistakes companies make when looking for software is not being clear on what they want. They get fixated on all the features platforms have, but don’t actually connect those features back to their needs.

You need to identify your key features early in your search. Sometimes you don’t know what’s out there, so a little searching has to be done before you can know what you value. That’s okay. But don’t get too deep into the process before you identify the key features you’re looking for.

If you’re looking to solve a singular problem, why look at full suites? But if you’re looking to solve multiple problems, a full suite of features under one roof might be the right answer.

Know what problem you’re solving and don’t create new problems along the way.

Ranking the features

Similar to the performance criteria, we want to rate the features in each solution a 1 to 5. We should only be rating items we actually care about and will use. A good bill pay module means nothing if you’re still going to use your old process.

When looking at the features you want, get really granular.

Don’t say “Bill Pay” and rate it on a scale of 1 to 5. We want to split it up into “Bill Pay approval flow,” “Payment processing,” and “Account coding.” These are just examples, but do what makes the most sense to you.

You won’t get all the granularity you need to start, so don’t be afraid to add items as you find functionalities you like. Then, rate each platform in that area.

Features to look for
So, you ask, what should I be looking for?

This list will not be all-inclusive, nor will it take into account your specific situation. I only provide it to outline my experience, as well as act as a “memory jogger” to help you as you set your roadmap.

1. Budgeting
2. Credit cards
3. Bill pay abilities
4. Expense management
5. Contract management
6. Vendor management
7. Document capture
8. Automation and AI
9. Reporting

Make sure you prioritize the ones that matter most to your business. Don’t let a software company sway because their bill pay is great when all you wanted was document capture.

It’s easy to imagine how you will use the software, but make sure you’re getting what you planned.

Lastly, be sure to account for the pain of multiple solutions versus the convenience of an all-in-one. All-in-one is nice, but if the system is not intuitive or easy to use you could end up ditching it completely. Make sure when looking at the all-in-one that it does a few of the core functions really well. It’s not worth sacrificing overall workflow for the ability to have it under one solution.

What is right for you?

One thing that’s often overlooked is who the platform’s target market is.

If the target market is venture-funded businesses, it probably isn’t right for a mom-and-pop shop.

With that in mind, let’s take a look at some of the different solutions.

Airbase

Airbase is best for small and medium businesses who are looking for spend management and no more. They offer procurement, accounts payable automation, expense management, and virtual cards.

They allow you to use your own card, but see Expensify for why that might not necessarily be good.

Ramp

Ramp focuses on domestic small to medium businesses. It’s also the only free solution of the bunch!

They promise quick approval and offer their own credit card that uses rules to make the user experience easier. They also offer employee reimbursement, spend programs, and bill pay.

This is a great solution for beginners that is easy to implement and integrates with many platforms.

Brex

Brex serves venture-backed startups as well as growing midmarket and enterprise organizations with a global footprint. That means remote employees, lots of travel, and multi-currency support. If you’re looking for global capabilities, here is a great article from Brex on how they’re differentiated in this area.

They’re also focused on AI and automation, which means they’re on the cutting edge of innovation.

If you’re midmarket or soon to be, check them out.

Divvy

Divvy focuses on small to medium businesses looking for a simple solution.

They use virtual and physical cards to allow users flexibility and auto-reconcile accounts to make the life of accounting easier.

Expensify

Expensify targets anyone with credit card spend and focuses solely on that solution. They allow you to keep the card you have and integrate it into their platform.

This sounds great but also makes for some complications. Transactions are slow to sync because of this, which doesn’t allow employees to finish out their charges as quickly as with some other solutions that use their own cards.

SAP Concur

SAP Concur targets enterprises and large clients. They’re one of the largest and most robust platforms, but you’ll pay for it.

If you’re looking for an all-in-one, they do it all.

Wrapping up

When doing research, people often overlook case studies done by companies. Obviously, we know they’ll paint a rosy picture. But by reading these, you can learn more about the capabilities and specific use cases. When you see a company similar to yourself, this should be a good sign.

For example, Brex has a case study about Medicinal Genomics using it alongside NetSuite. If you use NetSuite, this gives you great insight into how they’re integrating.

To wrap up, being thorough in this process prevents you from making mistakes that cost you money. Remember: time is money. Mess up the selection process and you’ll spend a lot more time finding another one.

This article originally appeared on Frameworks & Finance.