Distil Networks is built around a single but crucially important concept – preventing ill-intentioned “bots” from inserting themselves into commercial websites and stealing or otherwise misusing proprietary content. One of the ways we do this is through a global content delivery network that relies on Internet connectivity between 17 global data centers and multiple Internet Service Providers (ISPs).
So, as co-founder and CTO, my life has revolved around building and ensuring the reliability of our global network since the company’s founding in 2011. Along the way, as you might imagine, I’ve learned a few tricks that I’m happy to share with you here, in the hopes it will save you sleepless nights and unnecessary expense in the future.
1. Don’t overbuy
Right from the start, you need to understand the capacity you're planning for and the triggers for moving to the next expansion phase. We started out with six servers on Amazon Web Services, billed hourly and paid for by credit card.
2. Utilize Credit to manage cash flow
Using business credit for evening cash flow is a great way to even out spend spikes during the early cash crunch days of getting started. Having a monthly pricing agreement means you essentially get to have 30-60 days of float, this can be critical in preserving cash as you’re trying to finalize a round.
3. Expand when you’re ready
Even if you’re managing credit and float brilliantly, over expanding will quickly drain cash reserves. This is where cloud services are extremely useful, before we got into TechStars and raised a series A, we only utilized scalable cloud resources so that our utilization was paired 1:1 with demand.
4. Always pay your bills on time
This is a clear no brainer, but… early on we switched credit card companies and forgot to update one of our vendors which led to us missing one service payment. Boom! Our service was disabled and our customers were unprotected for several hours, logging the only total outage in Distil history. We recovered and made the determination we were ready to fly solo with Internap as our first direct vendor contract.
5. Trust is often more important than price in a vendor relationship
No matter how special an end-of-quarter deal sounds, it'll still be there at the beginning of the next quarter. When you’re in a service business, as we are, you need to know your infrastructure vendors have your back (see #2 above). Investing time in building a trust relationship with your vendor is far more important than a one-off price deal when you're in it for the long haul.
6. Don’t mortgage the future of the business for a great deal by over committing
When you’re starting out, service providers will do their best to press you into a three- or even five-year deal in exchange for bigger discounts. Don’t do it. Prices are getting cheaper all the time, and you’ll end up tied to higher payments than your competitors. If you can do a 12-month deal, that’s optimal, but never buy further out than 24 months max.
7. Make your vendors your partners
But don’t let reality get in the way - feel free to promise more business than you are initially able to commit to in order to get a better deal (but keep it believable!). The one thing I really wish we’d done and which we didn’t was to build our own test lab, so we’d have benchmarks to measure vendor performance. It’s crucial that you monitor third-party network performance before you sign any contracts, because your whole business depends on being able to deliver on your customer SLAs. We made the mistake of skipping that step, and it cost us a lot of money.
8. Don’t put all your eggs in one basket
A hybrid approach of self-managed and outsourced infrastructure will give you a higher degree of flexibility, is cheaper than going 100% in either direction, and gives you a failsafe if (when) something goes wrong. Outsource the faraway stuff – it will save time, travel expenses – and your sanity. Using both open-source and commercial software also helps to spread the load.
As our business has grown, and our ability to finance infrastructure expansion has solidified, both through customer revenue and additional outside investment, we’ve continued to build on many of these same principles. As we continue to expand our global network footprint and roll out Distil-built services, I see no reason to change our approach. And we still insist on running our own performance tests!
About the AuthorMore Content by Engin Akyol