Hacker News Reader

Comment by mikeryan | Hacker News Reader

Strangely enough my biggest concern is how well this will scale.

EDIT: I'm not sure why I'm getting down boated. I'm pretty serious. I have no idea the average task duration but assume its an hour then every "exec" can only do about 7-8 a day (assume some non billable time). If they get 10,000 customers and say each need one task every 2 weeks. Thats 5000 tasks and would require a staff of 125 execs. With 125 execs now you need managers and quality control and customer support. I'm assume execs are only getting paid about $15 an hour. That means A. Quality of Execs would be an issue. B. You only have about $10 of every hour left towards margin. Margin in straight services is nasty and this one is thin and in this case unless you have an extremely efficient staff your margin tends to go down as you grow - not up.

View on HN

Your managers are algorithms.

Your onboarding costs you $20 for each Exec you sign up - in this economy, your acquisition cost for them is likely near zero because WOM and free PR will give you more opportunities than you can handle.

You bootstrap CS with either the founders or one of your close friends, building systems to support the common cases (e.g. "Exec flaked on me") and using T1 support as a system-level T2. Your margins are asymptotically the same as selling iPhone apps at $15 each ($10 minus pocket change either way), but you'll have many customer LTVs in the hundreds. Eventually, you'll have a proper CS team in place like e.g. Airbnb and the'll involve themselves in < 5% of orders at a fully-loaded cost of $12 per incident.

We know that this model in broad strokes works because Rentacoder, ODesk, 99Designs, Airbnb, etc etc all work. The major gamble is that there exists sufficient market demand among the upper middle class for semi-interchangeable gophers. It doesn't strike me as obviously a billion dollar business but then again I thought "Young ladies will happily pay money to stay the night at private residences" sounded categorically insane the first time I heard it and we already know how that story turns out.

P.S. Numbers here come from "Big dumb megacorp with a traditional CS department assumes an American FTE talking on he phone costs $12." There's all sort of ways you can cheat that assumption, beginning with not being a big dumb megacorp.

Your managers are algorithms. - We know that this model in broad strokes works because Rentacoder, ODesk, 99Designs, Airbnb, etc etc all work

I'm sorry I don't buy this, I may be misreading how Exec works but it looks like Exec's are employees - not freelancers. I don't know how this would work any other way you couldn't guarantee staffing levels and "real time service" for your customers if they were independents like Task Rabbit. Not to mention Rentacoder, Odesk and 99designs all usually have some aspect of geographic arbitrage which brings down their prices. The major difference in this model is that Exec becomes entirely responsible for QC - unlike each of the other's you mention where if a person (or place) is crap that "vendor" gets downvoted out of the game. Here the only people who lose karma is Exec itself.

This brings a whole world of pain and management which none of the crowdsourced models address - payroll taxes, labor codes HR - scheduling. You're definitely not managing with algorithms. I have a good friend who runs a Doggy Day Care in SF who hires a lot of folks in this price range. He has a hard time getting 15 people who can walk dogs with a decent level of service.

You guarantee request quality of service, not staffing levels, by using a per-request queue. Beginning with the person most likely to say yes desirably, send out SMSes saying "Exec! New job for you. If you want details reply in next 2 minutes.". If they reply, opportunistically lock job for 5 min while they read. If not, move onto next potential exec. Tune parameters as required until 90% of customers get called within 10 minutes.

I totally get that your friend deals with the problems of flakey and unreliable people who are habitually not up to the task of working... but she doesn't deal with them like they were IP packets, does she? If she did, she wouldn't really be in the dog walking business anymore, but she'd also have potentially quite good reliability systemwide based on a network of high-failure components.

There seems to be a disconnect on the points we're arguing here. Task scheduling and routing is (almost) the easy part or at least a pretty solvable equation - there's a whole range of operations and queue theory texts that can make a quick go of managing that.

I'm more concerned with the human element. HR overhead, people making sure tasks are getting done well and in a timely manner. "but she doesn't deal with them like they were IP packets, does she?" - No she (he actually ;-) can't they're people.

"I'm more concerned with the human element. HR overhead, people making sure tasks are getting done well and in a timely manner. "but she doesn't deal with them like they were IP packets, does she?" - No she (he actually ;-) can't they're people."

That's the point of Exec and of every we-can-disintermediate-this-or-make-it-massively-easier-by-using-IT-and-CS startup. They can treat people like IP packets. At some stage they'll have practically open hiring and most people will wash out fairly quickly but the ones who either excel or consistently fail to fuck up will just keep on getting tasks from Exec, the others won't.

Think he's talking about the employee/customer relation and its logistics. Those sites mentioned are mere platforms or markets, it's up the both parties to find each other – they just make it easier by being the middelmen.

Exec is actually representing the seller side, the service provider. Execs don't match sellers and buyers, they match buyers with Exec employees which, already mentioned, can open a new set of challenges than sites like Airbnb and 99designs.

You definitely get more quality dealing with Exec's handpicked employees, but from what I see, there's alot of managing effort required from Exec unless they can (or have already) come up a way to answer it via algorithms.

This is the kind of thing which scales pretty well -- there are lots of unemployed/underemployed people in most markets.

I think the difference vs. taskrabbit is that it being realtime requires the task-runners/execs to be higher quality and more committed, so it might not work as well in a market with 0% unemployment where everyone has a 9-5 job. I don't think that's a major risk for the next decade or two.

They can obviously raise pricing. At some point, people will pay based on the money they save, and runners might get really efficient at certain kinds of missions. If, for instance, a common mission is "pick up my car from the bar the next morning", the requestor might think it is a 2 hour task (bus or taxi, ...); maybe a task-runner has an efficient solution involving a folding bicycle, or two people. Worth $100 to me, costs 30 minutes of someone else's time who is happy with $25/hr, win/win at $50.

People relate this to TaskRabbit and I believe that's wrong.

But not for the reason you might be thinking.

So why is it wrong to relate this to TaskRabbit? Because TaskRabbit isn't competitive, yet. They're about the auction format not the on-demand format. Apples and oranges. That's fine, but take stock in their assets: national brand recognition, solid network effects, a built out rep system, resources. It is only a matter of time on _if_ they want to compete.

If I were a TaskRabbit decision maker I would start a skunkworks team to clone, but not launch. Not until Exec does all the work of validating.

So really Exec is more relatable to Cherry.com, an on-demand mobile car washing service started by three valley insiders. I sign up, pay $30, go shopping, and get texted when my car is done being detailed. It's amazing.

And after talking to Cherry's contractors, there's also an amazing feedback system built up to ensure quality. But Cherry is just one vertical so it lacks scale.

That's where Exec could shine, by being the consigliere some of us need to get through modern life. Or while in a bind on the way to Burning Man.

On-demand dial-up services are the future. If you use Grindr, you know what I'm talking about.

It's closer to Uber than Taskrabbit, conceptually.

Reply on news.ycombinator.com