Seeking info/guidance on Internet startup costs

September 16, 2007 12:56 PM

We’re seeking advice on startup and ongoing costs specific to an Internet venture. We would really appreciate any advice/guidance at all.

Here is some background info…My wife and I have MBAs (marketing and finance, respectively) and have worked successfully for several years in our respective professions. We also have our own seed capital of $200-$300k for our Internet venture. Neither of us, however, have a technology background. (We do though have in mind an MBA acquaintance that is in the high tech sector that we would approach if this idea turns out to have any legs.) At the outset, we will work from our home and my wife will be working on it full-time while I will be working on it part-time. Ok, our Internet venture idea amounts to a… Yelp-like user-generated directory and social networking platform….surprise!! Please keep reading though… Speaking in general terms, our site will service an overlooked target market and start off by just focusing in our relatively small city (roughly 1 MM population) for now as we work out the kinks. Our city has a particular target market for our Internet site of 50,000-100,000 people and at least 500-1,000 small businesses (or representative of up to $200 MM in sales). Revenues will be derived, by and large, from general advertising and sponsorships and we’ll likely use Google AdSense initially but then possibly move on to direct relationships with advertisers. If we were to scale the system for nationwide penetration….then the target market could be at least 8 –15 million individuals and close to 1 million relevant small businesses (or representative of at least $150 billion in sales). Now, while we’re aware of some of the basic startup costs involved (e.g., computer/printer/fax, incorporation fees, trademark registration; logo design), we’re particularly interested in learning more about large technology-specific costs associated with starting an Internet venture these days… such as the Web design/buildout cost at the startup phase and then dedicated Web hosting costs (and buy/rent servers at some point?) on an ongoing operating basis. Those really appear to be the largest tech-specific costs. Moreover, for security, control and regular meeting reasons, our preference is to use a local Web design shop (we’re not at a major metropolitan city and so costs shouldn’t be that inflated)…rather than outsourcing the creation of our site to a freelance designer here or abroad. (For the creation of a simple Yelp-like system…would there really be that much of a cost differential between a local design shop versus online outsourcing?) We’ll, of course, seek to ultimately get quotes for the build-out/creation of this system from several local Web design shops…but it would be great to get some initial sense on these cost levels. If folks could possibly help us then in identifying and quantifying Internet-specific startup/ongoing costs (i.e., educated/seasoned guesses on the range of dollar costs or unit metrics or “rule of thumbs” involved)…then that would be extremely helpful for us. For example, with Web design/buildout, it be useful to hear what the possible various sub-cost components are to such a buildout of a Yelp-like systems; and it would be equally informative to get a sense of how much Web hosting will cost us (directly or via the local Web design shop?) if we make certain assumptions on how many people start visiting the site. If there is also any other advice for this startup process (e.g., what questions to ask the design shop; how best to draft a schematic/layout; what more I need to present to the Web design shop; etc.)…we would very much welcome it. FYI, our +50 page business plan is nearly complete, we have a domain name and we’re now trying to finish a draft of the site design details/layout/schematics. Thanks in advance for any help.

posted by boone to computers & internet (19 comments total)

uh... what was the question again?
posted by wfrgms at 12:59 PM on September 16, 2007 [1 favorite has favorites]

I think the question is, "Other than the usual costs associated with startning any business, what are the costs associated with starting up a mid-size commercial website, and how do we estimate them?"

I'm not knowledgeable enough to answer the question; some things you'll need to figure out, possibly with the help of the web dev shop: bandwidth costs (both average, GB/month and peak, Mb/sec); backend/computational costs (server(s) for running whatever databases or scripts the site depends on); initial development costs for the software; ongoing maintenance of the software; administration of the website (who does the backups? who's on call for 4AM reboots?); how much do you want to pay for uptime guarantees.

A lot of these costs depend sensitively on implementation details of the website, but I assume that someone with more recent experience than mine could come up with ballpark figures.
posted by hattifattener at 1:37 PM on September 16, 2007

Also it might be a good idea to hire a designer. It's hard to tell how important a simplistic and beautiful layout really is, but from your post alone (I was physically unable to read the whole thing, I'm sorry to say) I imagine the site you wish to start would benefit from a well-guided pair of e-scissors.

Empty space is critical, as if a consumer who can only see a few main ideas at once and instantly recognizes them is more valuable then a potential customer who sees an enormous WALL O' TEXT and gives up. I don't know what that would cost, but you could probably find quotes pretty quickly with google.

If you want specific figures as to server costs and such figures, you might want to study and contact some owners of other startups. Y-combinator (ycombinator.com) is the only organization I can think of that funds a whole bunch of different companies every year and (more importantly) lets them network and interact with each other, but there's probably more and possibly better resources elsewhere.

People there would probably be pretty knowledgeable.

Also: Paragraphs=awesome. Seriously.
posted by sandswipe at 2:02 PM on September 16, 2007

Not having an MBA going off of hattifattener's distillation, it may be useful to think of online hosting as office space that can be increased/enlarged at any time. The flip side of this is that the technical people you hire are somewhat tantamount to having your own general contractor onsite to do make the additions to your "officespace."

I'm not sure if this feeds your needs at all, but most of the question kind of seems like "MBA + idea + domain name + ? = Profit!" so maybe it's the implementation details that you're missing. You'll need something of a prototype or proof-of-concept to start with, so maybe you use the person who comes up with this as your lead programmer, or you make them your technical partner/CTO or possibly co-founder, depending on how much you trust/like them. The implementation will need to be managed and directed.
posted by rhizome at 2:17 PM on September 16, 2007

Ok..ok... I got a little ahead of myself. Thanks for responses thus far...it's helpful.

By the way, hattifattener, thanks for converting my Executive Summary-style query into an actual simple question.
posted by boone at 2:36 PM on September 16, 2007

boone, not to pile on... but your comment about the original "question" being an Executive Summary Style query is alarming.

Break all that stuff down.
posted by FlamingBore at 3:00 PM on September 16, 2007

Start out on a common platform, generally LAMP will suit. Choose a portable (multi platform), modular CMS so you can transition without too much effort to a larger platform. Start modular so the pieces can be broken off to dedicated resources as the infrastructure is required to scale.

Use dedicated hardware, don't do shared. Find a geek willing to build this for you and listen to them.. If you don't build it and go to an all in one shop, still find the geek and have them vette the other places work, and by geek I mean someone who has launched sites, infrastructure/etc and has been doing in for at least 8 years consistently. He/she should be familiar with whatever platform is built/launched.

Figure 10k in hardware for 3 decent servers (dual power supply etc etc), half rack in a tier 1 provider with 10mbit roughly 1k/month (haven't checked, guessing, the stuff I build is on order of 20+racks and we cost differently).

Your servers would be web, app/db, dev or some combination of the above. Speak with your hosting provider about providing firewall services or take 2k of the above and build a BSD or linux firewall and get the hosting provider to provide multiple isolated vlans. If you want to go full dedicated get a used Cisco switch (3750G or 3750 10/100) and a used Cisco ASA5510, roughly another 10k in hardware.

About 20k max in hardware to start which in a bare bones fashion at the network level should be able to drive about 100mbit sustained without issue.

If you do go shared hosting the metrix change all together, I don't know much about shared hosting other than I avoid it. It is, however, much cheaper.

Designing a high throughput, high performing and highly available system is a big exercise and expensive, keep hardware under 10k and go cheap with the bandwidth until you need it.

Your largest cost for the first year should be the person running it unless you get crazy return and hits, in which case both costs should rise in a pretty healthy manner.

There is a large amount of design+scaling factors you're going to need to evaluate on a monthly basis as the business grows or gets stale or deals with traffic spikes.

Form a solid relationship with a local used gear dealer who will warranty the gear you purchase from them.

Negotiate HARD with the hosting provider for SLA's, you're not going to be able to afford 4 nines (99.99% uptime) and not going to have the technical expertise to form them on to it. Aim for 99.9 or 99.5 uptime SLA's on your internet links, power and systems (if they manage them).

Negotiate hard with your vendor and lay out expectations, ownership of tasks as well as IP before selecting them to host, include outs for yourself if they do not meet SLA's, do not respond to trouble reports or do not provide the ability to expand your footprint or contract your footprint on a negotiated basis, 6-12 months recurring.

Find someone who has done this before, a third party who is not responsible for the day to day and have them vette the contracts, the technology decisions and the assumptions associated with the technology. Don't give them equity, hire them to find problems and get your hired or equity technology person to defend or justify them.

Require your technology resource to document the hell out of everything and to keep a functional "runbook" describing how the infrastructure works in case the technology resource bails. Have the person above vette and check the runbook procedures.

A deal of the above can be ignored if your design shop is also responsible for the operation and you have a decent contract. If you go to a single source for the design, launch and operation, insure critical things like the runbook, and all documentation/IP remain your property.

Good luck!
posted by iamabot at 3:53 PM on September 16, 2007 [2 favorites has favorites]

Lets see... we have two MBAs with no technical background (to the point of not formatting text to community standards, then calling it a summary). There is a 50 page business plan, but no elevator statement. Nobody in the bios list has the technical chops to whip out a prototype -- and they're looking to start a bootstrapped internet business venture on a six figure budget?

Please don't.

If you would like to put seed capital into an internet venture, contact your local band of angels (or contact me offline, and I can put you in touch with people who have much better ideas and track records).

You see... you don't know what it is that you don't know, and that's pretty much a recipe for disaster.

I'm really not trying to be nasty here, but to build a successful internet property from zero, you need to be someone who has at least worked on an internet property in the past -- and has at least a good sense of what it takes to run one.

I'm not saying that founders need to be programmers or designers or hackers, but founders at the least need to be able to evaluate those sorts of people, so they can find one who is both superbly competent and who shares a passion for the subject matter. And you absolutely need both of those attributes in your first technical employee (or third founder).

You should already know that your leading technical costs are not your servers (100k users on a LAMP (or similar) setup for a yelp-like system, can be done for less than $5,000 a year in leased hardware with gobs of bandwith -- look at self-managed dedicate UNIX servers on tier 1 bandwidth).

Your most expensive technical expense is your technical personnel. Without them, you have an unoriginal idea -- nothing more. Or did you think you were going to hire someone to wave their hands over some code, and conjure up a (profitable) site that runs itself without the care and feeding of a proficient individual (if not an expert).

You wouldn't be asking "I think I'm going to open an auto repair shop, but I know nothing about cars (though I do drive one daily). Since my hydraulic lifts are going to be my largest cost, can you tell me how much I should expect to pay for enough lifts to fix 10 cars a day?". No... you'd be asking about how to go about finding or evaluating a good lead mechanic -- and then he'd be able to tell you everything you need to know about the lifts that you'll be buying.

If you just want to run a social network, learn how to use ning.
posted by toxic at 4:56 PM on September 16, 2007 [4 favorites has favorites]

The people around here can tell you how to do it, many of us do it professionally, but on further reading, toxic offers some sound advice.

If you're a fresh MBA your program probably has or had a mentor ot steward who can get you connected to the right people who can offer technical evaluations.

To put something in to perspective, as far as how expensive technical folks are, someone with the experience to launch what you have above based on the data you've provided, keep it running and know what to do when it's broken can command 160k *easily* in the market place with a healthy bonus and benefits.

With that said, it's not unreasonable to do this, the key as you can tell, is to find the right technical person who will be candid with you about the feasibility of what you're trying to do.
posted by iamabot at 5:14 PM on September 16, 2007

Tough crowd...but good advice and info all-around. Thanks
posted by boone at 6:50 PM on September 16, 2007

We're the easy crowd, trust me. The tough crowd takes your IP or your cash and leaves you with contracts that eat your soul and kick your dog.
posted by iamabot at 8:15 PM on September 16, 2007 [2 favorites has favorites]

Let me get this straight... you want to start an Internet company, but know nothing about technology? That's not a very good start. Why not start a carpentry business, or a engineering company?

It's probably not a good idea to start a company where you won't even know how your company works, except in the vaguest sense. The Internet may be "hot" but there's plenty of other businesses you could start that would better utilize your skills. Alternatively, find some really great tech people and become angel investors or partners with them.
posted by lubujackson at 8:38 PM on September 16, 2007

I cannot urge you strongly enough to work in the industry before starting this venture, even it it means taking a $90k/year paycut and being a gofer at some dumb start up in the closest "hip" city. You are going to lose all of your money and you won't even know why.

If you absolutely insist on doing this, then I am begging you to hire a part-time professional copy editor to review and correct your outgoing communications. Your writing style does not engender good will. We are telling you these awful things because no one wins if you fail - except for the people who will take advantage of you.
posted by Optimus Chyme at 8:54 PM on September 16, 2007 [1 favorite has favorites]

We are telling you these awful things because no one wins if you fail - except for the people who will take advantage of you.

As one of the tougher members of what is actually a pretty friendly and candid crowd, I feel it's important to reiterate that, for its wisdom is deep.

Despite the fact that Internet professionals dress like slackers, and make what seems like astronomical sums of money to sit around and type a few lines each day, in between readings of fark, metafilter and slashdot -- it is not at all easy to run a medium or larger web property (though those who are good at it, make it look easy). There are lots of hustlers who will gladly take your money, while delivering crap and telling you whatever you want to hear. If we're a tough crowd to you, be very wary of any crowd that seems un-tough.

That symbol of marketing gone overboard, the Pets.com sock puppet set the industry back at least five years. I'm telling you to rethink your plans, not because I'm an asshole, and not just because I don't want you to lose your savings, but because I care for the health of my industry -- a poorly executed marketing-centric Internet startup perpetrated by people who have no business being in the industry hurts all of us.

Don't be a poseur, come join us in this industry. But do it the right way, don't plan to start at the top. If you want to get into the Internet startup business, go take a job with a tiny internet company that has an idea that you can believe in -- even if this means low pay in another city.

Find out what makes it all work, and why it isn't black magic (and why some of it is). Discover why your sysadmin is always disgruntled, and find out why his compensation package seems so high to you and low to him. Learn to spot the gleam of passion(*) in the eyes of founders and true believers, and understand that not only that it can't be faked, but that no startup becomes a midsized business without it.

Then go start something that blows us all away.

(*: you can see it in the corners of the mouth, while staring into the backs of the eyes. It helps if your own expression mirrors it.)
posted by toxic at 10:48 PM on September 16, 2007 [1 favorite has favorites]

I know they are harsh, but there is much wisdom in the above answers. The hostility seems to stem from your assumption that the technical issues of designing and building a popular website are a commodity that can be farmed out. (Like electricity.) Webhosting itself is a commodity. Web development is not.

What the posters are generally saying is, "You can just hire some webdev company, but they will take all your money and leave you with exactly what you asked for, but mostly likely won't be what you want or need."

I know you have a 50+ page outline of exactly what you want them to build - but trust me - those specs will change. You will realize that some ideas were dumb, some need tweaking, and you'll come up with new ideas. The devil is in the details. Design and development is an iterative process. The more iterations, the better.

To get the webdev you need, you need them to have some skin in the game. Why spend a weekend adding a feature, or tweaking some tiny, miniscule part of the site. Because it's their baby. Because they want it to succeed. Because they want it to be perfect. You are not going to get that with a project fee or even hourly rates. You will get that with bonuses, equity or whatever.

So forget about webhosting costs. They are trivial. Focus on incentives. How to attract and motivate the technical talent you absolutely need to get this idea off the ground.
posted by kamelhoecker at 9:12 AM on September 17, 2007

I don't think you're necessarily doomed, or should undergo years of training at another startup before doing your own thing, but I do think your lack of technical experience is an issue and you should approach the business differently.

You see, web projects are all about getting stuff built, and often it's impossible to tell if the idea is any good until it's up and running, yet you've tackled finance and a business plan as the priorities. You don't need a full-on business plan until there's a site capable of doing business. You don't need much finance until the site needs to cope with more traffic and features.

Find a talented, experienced web consultant/all-rounder and build a prototype for a few thousand. You'll learn an awful lot, and should then be able to launch something worthwhile with an overall spend in the low tens of thousands. Immediately you'll get useful data and feedback that'll let you adapt the site and your business model, and you can expand at a suitable rate.

It might not be as exciting as going straight in with grand plans, but it's less risky and more effective. I've seen non-technical people blow their savings on startups by ignoring good advice and trying to go straight in at the top, and it's not a pretty sight.
posted by malevolent at 10:15 AM on September 17, 2007 [1 favorite has favorites]

We appreciate the time folks have spent with their responses. There is a lot of constructive criticism and guidance here.

We continue to reflect on all that has been said. Prior to committing any hard dollars into this idea, we’d like to ask a two other questions as we examine our various options:

(1) Founding engineer: What would be an appropriate level of equity to offer (or partially sell) at the concept stage for a founding engineer?

We can create any set of ‘hockey stick’ projections that could help minimize the amount of equity being handed out. We don’t think that’s the best approach….but a certain level of base case projections would still appear to be needed.

Nevertheless, would a [20-50%] equity take that vests over a year or two OR upon certain milestones work? Any further thoughts on the equity stake and structure?

(2) Non-disclosure agreement: Going down the road of selecting a freelance designer or design shop… should we get an NDA signed up with them? Is this is a standard practice in this area? Or is it useless and a hindrance?

I already know that VCs don’t sign these….but our question relates to the hired tech person/team.

Thanks
posted by boone at 9:24 PM on September 17, 2007

(1) Founding engineer: What would be an appropriate level of equity to offer (or partially sell) at the concept stage for a founding engineer?

There are three founders. something just above 30% to each of them seems about right, with some dilution clauses to cover future outside investment (so when you take VC funding in the future, it dilutes all partners). Without this individual, your business is worth nothing. 66% of $100M is a lot more valuable than 100% of $5. (I'll include your mefi account as part of the business assets.)

Why do you believe that this individual is beneath you? This is not a rhetorical question... I'd like to hear the answer.

We can create any set of ‘hockey stick’ projections that could help minimize the amount of equity being handed out.

You really don't get it.

You're not interested in minimizing the equity that you're giving your PARTNER. You're looking for someone who shares your passion and your vision, and who wants to make your business work -- it's just that they have the technical background, and can actually implement your grand ideas.

You're not "creating" anything for him, especially something with an agenda. You're telling him the honest truth, always, even when you think it's in an area outside of his expertise, or that he'd have no opinion on.

When he speaks up, he is your equal, and he is the equal of your spouse (if s/he is the third partner). If you're not prepared for that, then you're not prepared.

Offer to cut him/her in under the same terms that you are giving yourself. It's possible that they'd rather take more money and less equity, in which case, you've got a good start for negotiation.

As far as vesting goes -- if their shares vest, yours should too, but everyone's should fully vest immediately for a merger, IPO or other major corporate event (if the purchasing company wants to lock you in, they will through other means). Again -- you want to give this person plenty of equity, and you want to make them very wealthy -- because that equity will make them want to make you very wealthy. Remember -- the stuff that they do is every bit as important as the stuff that you do, when it comes to creating value in the business. Otherwise, you're building pets.com.


If you are looking to minimize equity stakes to maximize your return, then just become a fucking investor already, and leave the building portion to the experts. I guarantee you that you'll make more as an Angel investor in Internet companies than you will by starting them -- even if you are a hot-shit Internet entrepreneur.

Either you're doing this because you love it (in which case, the money's not so important, and you'll be thrilled to give a third of the company to the right person, because the value of the three of you is much higher than the two of you alone), or you're doing this because you're looking to make some money (in which case, your odds are much better in an index fund)

"But wait", you say, "I can't tell who's worth 1/3 of my dreams".

The peanut gallery nods knowingly.


should we get an NDA signed

NDAs are mostly a formal pain in the ass. Most professionals won't disclose your business secrets, even without an NDA. Yet some of your embarrassing personal secrets, and most of your professional fuckups, will be shared when your employees talk shop with their colleagues from other, similar companies. Deal with it.

If you have a boilerplate NDA that's reasonable (think: less than one page long, less than one year duration), it rarely hurts to ask your interviewees/vendors to sign it -- but if they refuse or otherwise balk, then whisk it away and don't require that they sign it. You're not going to want to pay to enforce it, anyway. Also -- smart people have a tendency to question everything. Do you want to start off your relationship with the smartest girl in the room by presenting her with something that she might have a conflict with?

Besides... you're building a community ratings site for local-oriented personal service businesses, with a possibly more complicated backend (scheduling and referrals?) -- and you're hoping to get small business people (with service businesses) to pay a little extra to get featured placement, extra features, and for referrals. Do you really have Intellectual Property that's worth protecting?

(NB: It's fine to NDA your employees, contractors and other people for whose time you are paying. That's different from asking a job candidate. When you're offering someone something (like, say, a job) in exchange for their silence, it's a whole lot more acceptable than when you're just having a conversation with them.)

Oh... and entrepreneurs who are openly starting a startup usually want to be contacted by interested parties. Did you mean to include an email address on your profile page?
posted by toxic at 11:02 PM on September 17, 2007

If possible, I'd leave equity discussions until you're past the prototype stage and simply pay normal rates for that initial work; you need to make sure you're not wasting each other's time. But yeah, if you want them fully committed then make them an equal third partner.

The right person will spot several potential problems within a few minutes, offer new ideas, and probably be quite harsh about your prospects yet enthusiastic about the possibilities. You need hard questions and strong views, so don't fall for the trap of being charmed by someone who tells you everything is great.

toxic's right about NDAs; by all means have them, but don't waste time on excessive secrecy (is there really a lot to protect?) and don't be surprised if people won't sign them just to have an initial discussion (I wouldn't).
posted by malevolent at 1:00 AM on September 18, 2007