Manual Scaling

2013-04-23 by Jason Freedman
  • Vote on HNUpvote on Hacker News

At 42Floors, we’re trying to change the way people search for commercial real estate.  It’s an enormous undertaking.  Gathering commercial real estate listings data is not trivial.  The data comes from many different sources, all of which must be cleaned and normalized.  We sign agreements with landlords to scrape their websites.  We do deals with brokerages to take feeds.  We parse emails from industry distribution lists. And a half a dozen other messy methods.

So when you see a nice listing with a picture show up on 42Floors, it’s usually because of some fairly schleppy stuff we’ve done.   It would seem obvious that we have to automate all of these operations and that the technology behind routing all this data, normalizing it, cleaning it and presenting it, is the core technology of the company.  But what’s not obvious is just how long we’ve been able to go without building any of it.

Early on, we didn’t have website scrapers.  We just hired people to manually enter each of the listings.  As we started receiving inbound distribution lists by email, we didn’t immediately build our email parser to transcribe the email blasts.   We just hired some people off TaskRabbit to manually go through each email blast.

And as the numbers kept growing, we kept scaling our manual processes, we worked with an offshore firm to manually enter even more data. In fact, when you look at each aspect of our operations, you will find a fairly sophisticated manual operations process predating every piece of technology that we’ve built.  Our wonderful Director of Vibe, Alison, manages all these processes while she does a million other things.

 

So now, your turn.

Is manual scaling right for you?  3 Questions to Ask Yourself

 Have you achieved product-market fit?

If the answer is no, you’re a prime target for manual scaling.  In the pre-product-market fit stage, you are really just running an experiment.  Anything you do that allows you to test your thesis faster is worth it—even if it means doing a bunch of manual work that will have to be replaced later.

 

Would outsourcing save you time?

If you have a process in your company that could be outsourced, you should.  The cost of outsourced labor is so low compared to writing code, that you should almost always pursue it first.  Later on, when your startup is showing great traction metrics, you’ll have more time and more money to write the automated version that will be the future of your company.

 

Will your users even care?

If all you’re doing is changing some back end process that doesn’t actually touch your users, they won’t care.  They want their problem solved.  They don’t care if it’s through beautiful code or some hacked-together manual process.  If the latter is 10x faster to get to market, you should take it.

 

 

Find discussion of this post on Hacker News

About Jason Freedman

Entrepreneur, Co-Founder at 42Floors, Co-Founder at FlightCaster, YC-alum, and a Tuck MBA

  • Upvote on Hacker News