Faster WAS Better: 7 Reasons to Go Back and Do It Right Now

Why data integrations (and most other systems) should be revisited when startups become stable

Jessie Szemraj
Jul 27, 2017 · 5 min read

One of the most common mantras in the startup world is “done is better than perfect.” While this might be the best course for true startups, many companies apply this mantra well beyond those early days. Once companies reach more mature stages, the hacks we all threw together back in the day just don’t cut it anymore. In fact, they can be almost laughable in retrospect. For example, horror stories I’ve heard around town:

  • Using Gmail labels as a CRM system
  • Data warehousing on Dropbox
  • Managing thousands of segments via… Excel attachments

Particularly in the world of data integrations, these hasty decisions can have serious consequences in the longer term. The next time you need to make the case for spending the time to take a step back (aka a reliability cycle), use these motivators to make your case.

1. Missed opportunities

Are you dropping data without knowing? Did one design decision limit or even eliminate future products or business opportunities? Without revisiting the details you’re likely leaving money on the table. Lost deals and reduced revenue are the two most compelling reasons for most exec teams to dedicate valuable resources to revisiting past projects. While missed opportunities might be hard to measure, I’m sure your sales and customer success teams have an anecdote or two they can share.

2. Simplicity

Why have 5 ways of doing the same thing? Whether it’s understanding and explaining systems, troubleshooting, or product improvements, consolidating your hacked-together processes into one, more robust, process will provide lots of benefits in the long run. Don’t take this to mean you should have one and only one system; there are benefits to parallel systems for specific use cases. But take a look under the hood before dismissing this point, there’s probably lots of things you could consolidate without sacrificing performance.

3. Efficiency

Speaking of performance: you probably could be doing it faster. One system (or at least fewer) to maintain means you have more time to identify ways to make data move faster. Speed is key in today’s world (and a great sales tool!), so don’t let imperfect but working systems hold you back.

4. Reliability

Pissed off clients = lost revenue and bad press. Startup-era clients knew they were signing up for an… experimental… product. But the types of clients you’ll want in the long run will be less forgiving. There’s nothing worse than telling a client you’ll be missing that deadline after promising over and over data would be delivered on time. Save your support teams this pain and make sure your system can hold up it’s end of the bargain.

5. Scale

Don’t start from scratch every time. If each new integration requires fresh connections with unique requirements, you’ll be pulling your hair out in no time. The number of players in the data ecosystem alone means you’ll probably be working with many more than just a few companies in the long run. Make sure the framework you built out for just those first few partners will work for others down the line.

6. Documentation

Neglected but important. I bet when you first built that system in your “faster is better” phase, documentation got left behind. For a growing company, you can’t rely on an engineer or two to know everything and share it across the team. Not only is it impossible for any one person to know everything in a rapidly growing company, but that person is too valuable to be answering questions all day.

7. Transferability

Closely related to our last point, but important on it’s own. The original architect won’t always be around. Will new hires be able to figure out the cobwebs left behind? Documentation is definitely part of this, but even with detailed documentation, three different methods are harder to learn than one. If you’re building a system that can only be run by it’s original engineer, don’t plan on being in business for much longer.

What Now?

This is the exciting part. I know it sounds scary, but there’s lots of fun to be had in digging through the weeds and finding better ways to operate. At onNomatic, this is exactly what we’ve been focused on. So cheat a little and use what we’ve learned to kick off the project now.

Do a thorough audit. Challenge assumptions that things are working as intended. Run a test file through the complete process if you can and ask questions along the way. Internal stats are key here, and if anything, you might discover some great KPIs for future monitoring.

Use partner resources. Confirm stats with your counterparts. If you aren’t seeing (approximately) the same numbers, something might be off. Just remember to put each side’s assumptions out on the table first, because one doesn’t always equal one when it comes to data (uncookieable cookies anyone?).

Learn from scratch. Have a new hire or outside party like onNomatic dig in using resources and people across the company. A fresh set of eyes will ask the “dumb” questions, which might not turn out to be so dumb after all.

Write it all down. Not only do you want the documentation for future employees, but likely yourself. Writing the details down will force your team to dig into the weeds and really understand what you built, maybe even uncovering opportunities for improvement (or bugs!!) along the way. As an added bonus, you can use your internal documentation as the basis for external documentation/FAQs in the future. Trust me, your support teams will thank you.

Prepare for the worst. Issues are inevitable in the complex world of integrations. Standards, processes, and products change, and with change comes (hopefully) inadvertent error. Make sure you’re prepared to handle breaks when they happen and build a process to catch and fix them quickly.


onNomatic Consulting specializes in adtech data integrations. Whether you’re pushing data to a DSP, pulling from a DMP, buying 3P data for your clients or looking to sell your own. It’s a complicated world and we help you navigate and optimize it. Want to learn more? Check out our website or contact us now.

Thanks to Matej Horava for making sure I didn’t spell anything wrong :)

Thanks to Matej Horava

Jessie Szemraj

Written by

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade