No long distance teaching for New York City

You can't outsource education, says New York City. Socratic Learning, a company providing online tutoring services with teachers from India had bagged a $2 million a year contract to train 2000 children under a federal "No Child Left Behind" program.

New York City insisted on teachers being fingerprinted (which they did, in an Indian police station), and prints verified by the FBI, which was done.

But New York City wanted the teachers to be physically present, and for the company to provide them U.S. Social Security numbers, which aren't given to non-residents. Is this the end of tutor-outsourcing as we know it?

Maybe they can turn around and offer that to Indian students.

Time for the consultants?

When the Indian IT hiring scene is as bad as it is today, it's time to sit back and think. Why is it so difficult to find people?

Demand and supply, you say. There's few people who really understand coding and these people are hugely in demand. That demand has encouraged those that didn't really care enough about programming to enter the fray; and it might sound extremely arrogant of me to say, but we Indians seem to revel in doing things we don't really care deeply about.

From Arranged marriages [1]to Higher Education, Indians seem to prefer to "study whatever is fashionable for the job market", rather than following their heart. This hasn't just happened; our history has shown that the socialist days of the past meant limited jobs, and even more limited opportunities. A bad education could make one absolutely miserable in terms of money; trades like journalism, the fine arts or even sports were littered with lowly-paid degree holders who had learnt the hard way that there just wasn't enough in there. The next generation was obviously huddled into lower risk career options; an engineering degree, computer skills and programming lessons. They didn't dare to question their huddling, perhaps, and in reality, they bought the study-for-a-job argument.

Engineering degrees started to get a huge demand for Computer Science and Electronics degrees, and those in the rest of the streams like Mechanical and Chemical Engineering took extra classes to learn C++ and Java. Not from a burning desire to learn, but to make the grade for the Computer jobs.

With the advent of the tech outsourcing boom, companies started to look for people and initially, the Electronics and comp sc. students weren't enough. So they would hire engineers from other branches, who went through a rigorous training program and learnt the ropes. Everyone who knew anything about computers qualified.

Soon, even the engineers weren't enough so companies were hiring among any graduates of any school that had seen a computer at a distance less than 5KM, though this requirement could be further relaxed if a project was slipping past.

Switch over to now. There are a few good men and women. And there are a huge number of qualified but unenthusiastic engineers in jobs that don't suit their temperament, but who's looking at temperament? This is a science, not an art, they all say, and slap you with the 10,000 page ISO900x manual and the CMM process document. Anyone can do this, if we break it down.

But can we? Is software development so simple and definable that we can write a manual and expect that a team, ANY team, will deliver to expectations? Not one single manager worth his or her salt believes that. It's eventually about people, and about having the right people. The code gurus are the real people we need, those that can take a requirement to its logically intended software end. Not necessarily the superheroes, though they are preferable to the "junta" crowd, but "developers".

So where are these code gurus? Some are managers now, since companies don't allow you to evolve much or earn more as developers. Some get disgusted with the sheer lack of quality work and either turn into entrepreneurs or join marketing or sales to beat the boredom and learn something. The rest are treated very highly by their companies and manage their projects fairly well.

There's a problem there, though. These guys, though useful, aren't well paid. Because the Great Indian Software Dream has only one career path - managerial. If you want more money, you become a manager. You're defined by the number of people you lead, which is such an irrelevant measure I splutter everytime I'm questioned why I hate it so much. Good coders often don't give a damn; they want to build something, not someone. But if money calls, one has to go up the manager route, does one not?

Maybe it's time for them to turn consultants. Then they can demand higher pay (per hour) than they usually get, work for short periods (3--6months) and spread over multiple companies. They'll not have to turn managers to get paid more, and so can continue to do what they love most : code. Lesser office politics, and perhaps less red tape to deal with. Delivery is all they're hired for, and that's all they get.

Yes, startups will want these guys. But startups should understand the demand-supply equation - don't let a person go because they're not willing to work full time. Heck, even pay them partly in stock (yes, you can still do that in India) if it makes them want to stay.

It might just be the age of the consultant quicksilver software developer. I think I'd like that as a career sometime :)

I'm just afraid of what happens if the consultant market gets muddy. If hiring a consultant is as tough as hiring full-time, you'd always prefer the latter. Enter blogging and networking: only the fittest among the networkers/bloggers will survive. And Googleable-names will be the top consultants.

1. Arranged marriages aren't always decisions taken by parents for their children; in fact most counter arguments claim that every individual has a choice to say "yes" or "no", in all but the most remissive clans in India. Yet, the concept remains that parents provide the choice(s), individuals only approve whoever they like. There's a huge difference between attempting to find someone (and most likely, finding yourself in the process) and having someone else do that for you. Oh, it is changing, by the day. But it speaks tomes when you find those that didn't even care enough to try and find the person they will live with for the rest of their lives.

Changing names (and the World)

Guy Kawasaki has changed the name on his blog to "How to Change the World". That, I like. Also I like the tagline:

"A practical blog for impractical people."

Only impractical people can change the world. Practical means, or so I think, "connected to the world", and by definition, changing the world involves reversing practicality, if such a word exists.

Have I ever thought of changing the world? When I was fifteen there was no other option, for there was nothing that deserved to stay. I wanted to change the way education was slapped upon us, I wanted to make cars cheaper, Buying things for indulgence would be the norm, I'd make a world class hospital, and change politics for what I know.

Now when I'm 32, I don't see it as black and white anymore. A hospital's fine but I want to be able to pay for private rooms. I want to change politics but not right now. I can change education, but the effort needs me to go full time, and I have a family to feed. I want to make cars costlier, except the one I like. I want to keep that shade of grey, and can we remove the 50% gray, please?. Oh yes and that entire series of grays has to go, but that one out there, peeping out from the corner, yeah that can stay. I kinda like that.

I kinda like that.

Famous last words.

"Getting Real" is now free

The folks at 37Signals.com have made their Getting Real book free to read - read it on the web now for no charge. You can also choose to buy this book in PDF or paperback form.

This book is an awesome read. You may not agree with everything you read (I didn't) but you have to admit that it makes a fascinating argument for a new world - the New Web Startup. (I hate the term Web 2.0)

What it brings to the table is a collection of startup concepts you might have heard of in passing that contradict popular notion, and somehow the authors come across as the Robin Hoods of startups. You've heard "Get a VC", they say "Screw the VC, get started anyhow". You've heard "You gotta go full time", they say "Part time's ok, even weekend time". "More features than the competition" is now "Less but better tested features".

There are small, tiny little things in there that stick on for a long, long time after you've read them. I find myself sitting back and thinking again about some of the assumptions I'd made, and asking myself, "What was I smoking?" For instance, you'll find the argument against functional specs compelling:

Build, don't write. If you need to explain something, try mocking it up and prototyping it rather than writing a longwinded document. An actual interface or prototype is on its way to becoming a real product. A piece of paper, on the other hand, is only on its way to the garbage can.

What you might want to is get yourself a good 2 hour undisturbed stretch to read this book. Maybe four hours, if you, like me, need to wander off and search for what struck you when you read something.

An imaginative workplace

The Volkwagen "glass factory" at Dresden is an incredible workplace, and the photos will simply blow you away. Here's a (linked) image:



I will (hopefully) be visiting the Toyota Factory in Bidadi (Karnataka) soon, and perhaps I can bring my photos here. I'm sure a clean and efficient workplace is possible anywhere in the world, and the VW factory simply blows you away.

Killing startups

Paul Graham talks about 18 mistakes that kill startups. When I read that first, I said: Why 18? Why not 20? Or top 10, or even top 3, in todays times of instant gratification? I think this was more on the lines of "Ok, here's the top 10. Oh shoot, I forgot that one. Let's make it the top 11. Er, remember that guy who shot his co-founder. Top 12. And 13. And 14...Okay 18 and that's it. The remaining 982 they learn for themselves."

Apart from what Paul said, I think one of the other BIG reasons that kills startups is:

19. Inertia
You know the standard way startups try to fund themselves: I'll do some consulting on the side, they think, and pay the rent. Soon, they get a few projects, hire a few employees on both product and consulting sides, and now they have some healthy revenue that they pump into their product's marketing like visiting trade fairs, buying advertisements and so on.

But the product still does not sell. The employees are working hard; as hard as they can. But some employees are on the consulting side and others, on the non-performing product. As every company does, founders try to reward performance through bonuses - consultants get the meat and the product employees are handed only a "company wide bonus"; they either begin to leave, or ask to be moved to the "other side". The products flounder further.

Yet, the startup owners are doing well - the consulting revenue is enough for profitability, and they still feel the product "has something in it". So they continue to fund it, and after a few more down or "just break even" quarters, have the big doo-dah meeting: What should we do with this thing? Founders are (usually) human beings with a huge attachment built up for the product; after all, it's the only thing the company owns as a technology asset, while the rest of the work is for-hire-for-customers.

Warning sign: Inertia.

They *could* fund the product forever. As long as consulting revenues come in, the product can underperform, or just about break even, and no one will feel compelled enough to break the impasse. This is where I believe startups fail; the inertia from running a business profitably eats into the potential growth one could have with a more "remove the deadwood" approach.

That's why founders need a plan. Not a plan you show investors where your numbers make you sound like a significant contributor to GDP. A plan that says: This is how much I want to grow. This is how long I'll wait to see that growth. And if we fall short, I will quit or find a venture capitalist.

I failed in that I did not have a plan. I did have dreams. I was scheduled for my second ferrari last year, according to one of them. I have one car by ferrari's parent company, Fiat, and I just cleared the loan last September. Oh so much not the same thing.

What I'd recommend is: 50% growth for the first five years of sales, and a Rs. 5 crore revenue target (Rs. 1.5 crore net profit) within 36 months of your first sale. Or quits.

If you already have venture capital, throw all of this away; the VCs will tell you what sound to make next.

Note: I understand inertia does not "kill" a startup - if you're still profitable, are you dead? Well, I could run a company profitably with $1000 in annual revenue. See what I mean?

There won't be no more VCs in India.

A Unified Theory of VC Suckage is a phenomenal article by Paul Graham about how the structure of Venture Capital funds explains why people who (wo)man them are usually jerks. It's the structure of the fund, he says, that pays them a percentage of assets managed, that results in their "suckage". Meaning they need more money to invest, to make more money as commission, and more money is not always good:

VCs don't invest $x million because that's the amount you need, but because that's the amount the structure of their business requires them to invest. Like steroids, these sudden huge investments can do more harm than good.

That's just lousy for the founders who need lesser money in todays world. Imagine you, as a startup, need some money. Not a heck of a lot, around a million bucks. Because you're either outsourcing, or have a lower-cost marketing idea, or you don't need as many servers as google, or simply that hardware and bandwidth costs are low enough to serve out of your basement. You're not going to get these piddly little sums from VCs, who want a minimum of $5 million per deal because they got a $2 billion base to work with and each guy has to deploy like $100 million and she can manage 20 deals at the absolute maximum.

So, the startup thinks: What the heck, let me get the money, at least more money is better than nothing. Well, it turns out it's worse. Getting more money means giving up more control, and now the VC owns you; they will then do the standard thing of putting their own CEO, CFO, C-everything-O until you're left with C-toilet-cleaning-O. That's still ok, I would guess, if you got paid for it - but Venture Capitalists shudder at paying founders for their stock.

So why am I ranting about this? Because this spells the death knell for traditional VC investments in India, which are already under some amount of duress.

1) Indian startups require a lot lesser capital than, say, a similar startup in the US. I don't mean a start up that sells in the US, I mean one that sells to Indian customers. A software startup with a great idea, 20 employees and a few customers and a reasonable growth plan will perhaps need about ten crores, which is about two million dollars. If VCs are coming with truckloads of cash, they need to back up those trucks a little bit and get them one truck at a time. Unfortunately their own structure denies them that opportunity.

2) Deal flow is practically non-existent. That means there are very few VCs whose ego is not inflamed by the idea of taking another VCs handouts. In fact, the little deal flow that exists is always an ACCRUAL; meaning a downline VC will pump in money into the company but refuse to buy out an earlier VC or angel; meaning there's a lot lesser enthusiasm for the early stagers.

Exits have to be through IPOs or acquisitions. Either option is perhaps too far away, and if founders aren't going to get some payout at the time of investment (part of their stock for instance) their ability and interest in keeping the business going is going to flag, especially if they are just C-toilet-cleaning-O. You will take away their control and management powers, and won't even pay something for what they do own (stock), and you expect them to work as enthusiastically as before. I may be quoting Bill Gates here, but that's the stupidest thing I've ever heard.

No wonder VCs are not getting the best deals; if founders know that VCs will take control AND screw it up (I'll talk about that later) they won't give up a good idea.

3) Okay, let's say we can fix deal flow by saying it'll happen tomorrow. Founders are still going to lose (and perhaps lose more control). If VCs don't help founders by giving them a little bit (equity or cash) on each deal, the residual founder ownership is going to be very low and they will walk away. VCs will be left with the business people that they CxO'd and those guys, to show they can deliver, will go about doing what they do best, making big-ticket deals. They'll just acquire each other left, right and center, with no real profit to talk about. Sounds familiar?

4) Valuations in the past have been unnervingly low. Take JobsAhead.com, a recruiting site which was sold in 2004 to Monster.com for about $9.5 million. They got (unknown amount of) angel funding from the Dalmia group and two rounds from ChrysCapital totalling $5 million. I'm guessing ChrysCapital wouldn't have got more than 70% of stock - that's a total return of around 60% over 3 to 4 years, about 12-15% annualised. Not exactly print-worthy.

Another example: Baazee.com was acquired by eBay for $50 million. VCs Funded: $22 million. Net return would be around the same - 12-15% or so.

More: NetKraft gets bought by Adea for $8 million. Actis & Jumpstartup: 67 percent paid $8 million. Net return: (MINUS 33%).

I don't know what happened in the above cases, but it seems to me like acts of desperation. Take the deal and run seems to be the mantra, and VCs have LOST money or face. In all the above cases, VCs controlled more than 50% of these companies, and I believe that they decided to exit while they were still profitable. Now you know why VCs become Vulture Capitalists!

Initially I thought founders would not have been too happy. But when you're C-toilet-cleaning-O, money for your stock is always good, so these founders would've been honestly happy. Some of them have gone on to start VC funds in India or incubators or become members of India's band of angels.

But I think these valuations kinda prove a point. This is what happens when you give VCs control - they sell out for stupid valuations, and they ruin it for the next set of entrepreneurs.

4) I think many VCs have figured out they can't run companies. VC funds are now coming as PE funds, meaning they invest in public companies and get stock in return; for in such companies there's a need for big money, and the funds don't get as much control, so they can't do things like change CEOs and CFOs etc. No matter what anyone says I don't believe VCs do this management changing rigmarole because they like it; they feel existing people don't have the pedigree or the experience they can bandy to THEIR (the VCs) investors.

The rules of the VC game are that you put in money and then you protect it by surrounding it by your people. The rules of the PE game are that you give the money and you get a cocktail + dinner presentation every quarter, and if you get all cranky you can go sell your shares in the market, thank you very much.

So is this the end for Venture funds in India? Note carefully that I said "traditional" VC funds. It's the end of the road for you guys. Pack up your bags and go home.

To the smart lot that's left; come hither and have a cup of chai. Let me be an arrogant asshole and tell you what I think. I think you guys belong here. This is where it will happen, where the biggest amount of money is. But it needs money, and you're the only ones that have it. You know this, but you're afraid of the past. The Indian past of low valuations, lousy VC performance etc. Forget the past. So some people screwed up. So what? You're smart. You're smarter than those guys boarding the plane back. You can find, and fund, and foster the next big thing here. Look at India, not just the IT sector - look at India.

Look at dishing out smaller chunks of money, and managing more excel sheets. Look at giving founders more breathing space in this non-credit-rated, low-debt-availability market. Look at larger valuations. Look at making money in 10 years instead of five; you're smart enough to convince your investors, I'm sure. And you know this already, but wait and take the IPO route rather than jump into a pool of choreographed acquisitions.

Your chai is getting cold.

P.S. Sramana Mitra has a neato-article on the same wavelength, calling for sub $25 million funds.

Eyesore 9000

If you're ever bogged down by standards like ISO 9001:2000, you must read the Eyesore 9000 standard. It comes with intuitive, easy-to-understand phrases like:

How you document it [your process] is your business.Feel free to use scratch and sniff in your Quality Manuals. Do the whole thing in Morse Code. This Clause specifically prohibits your registrar from mandating the format of your documentation, or the methods by which your execute ISO 9001. It's the Clause they don't want you to know about.

Such valuable advise, and for free! The e-book comes with a lot of handy suggestions like "Make sure the ink dried.", which has surely reduced the number of wet-ink accidents for the ISO 9001 implementers, formerly a major cause of implemenation related injuries. But you must have a little something the book expects you to have: a sense of humour.

Creating a software company: Processes?

Let's say you want to create a software "services" company. By that I mean a company that offers software development services to customers, perhaps on a turn-key basis. (As opposed to a product company or ISV) Now if you want to grow to a bigger employee base than, say 10 people, you will need some way to co-ordinate activities and keep things organised.

What you need, they'll all say, is a well defined process.

There are some big standards in this sort of setup: ISO 9000, CMM and the like. I'm not a fan of multi-letter acronym standards; for small companies, these are more a pain in the backside than of any serious help. And for the entrepreneur, these are an absolute nightmare!

What you want is effective project management, and the ability to learn from your mistakes. That's all. Here's my take on what you need, and how you can ensure that you can grow your staff strength and still keep your sanity. Note that some of this can also apply to product companies (ISVs), so if you're an ISV with some time on your hands (oxymoron?) I hope this will help you.

Ensure you have the following tools:
1) A stable version control system
2) A bug/defect tracking system, with post-mortem cause analysis (a custom combobox might do)
3) A time tracking system
4) A project management tool.

Then, start writing. You need to write down, CLEARLY:
1) How you specify work. If you work with fixed specs, ensure there's a place to store (and search for, usign google desktop or such) specs. If you work on T&A, ensure you put in feature requests into a common area (perhaps in the bug/feature tracking tool)

2) How you develop: This involves your entire software development process; write down the different ways you can. I.E. where the customer's involved, where dev, qa and systems come in, whether you use milestones and deliverables, when QA does regression, how you use automated tests + smoke tests on builds, how you use automated builds (please do) and so on.

3) Version control: What happens when a new project is started, what is the format of user checkin comments, do you use edit-and-merge or exclusive locks, do you expect personal workspace checkins or only buildable checkins (or both) etc. Work carefully with defects as well: bug fixes should perhaps be marked in some way with the defect code.

4) Defect reporting: Write about how defects will be reported (area, steps to reproduce, workarounds, "owner" of a bug, bug states, approval procedure etc.) And write about how you will analyse WHY a bug occured, as a post-mortem process after the bug. (bad Spec, bug reported wrong, dev error, feature-not-bug, etc.)

5) Time tracking: Write about how your people report time they spend. Is it per spec, per product, or freehand? Are you using billing codes? Can you create estimates and link them with actuals alongside?

Now try and link them all together. Run through a cycle of projects and come back and refine these processes.

You may find that you need:
a) Better Specification and feature control. There are requirements management tools that can help.
b) Estimation recording and analysis
c) Organised learning and a central knowledge base

Or more such things. Now schedule a review of your original documents after every project cycle, and you now have a process of "improvement" also.

As a uISV, you probably hate this kind of stuff. But if you plan to be more than a handful of employees, having this planned out is essential.

Over the next few posts I'll try to write a bit about each of these topics, and a sample process document for each of the above.

Free Tools for Micro-ISVs

A discussion on Joel's business forum showcases free tools for Micro-ISVs (a term for ultra-small Software Vendors, usually those that have another day job that pays the bills).

Don't miss it.

How much should you charge as a consultant?

A post on the JoelOnSoftware Business forum wonders what one should charge as a consultant, and whether dividing required income by 1000 is a good way to set an hourly charge.

Typically, as an employee, you would make some money as salary and some as benefits. Benefits include employer payments into your retirment or pension plans, medical insurance, gratuity etc. Usually this is around 25-30% of your base salary. So if you made Rs. 10,00,000 (10 lakhs) a year, what do you need to charge if you become a consultant? (In India, most consultants would charge by the day or the month so I'll stick with that charge)

What's different between an employee and a consulant? Lower benefits, yes. So you've gotta make that up somehow - 25-30% of base must be added to the amount you should make.

But we haven't considered taxes. If you are an employee you get a few deductions for housing and that's it. For a consultant, you can expense out a lot more - the depreciation on your car, the cost of fuel to get to work, what it costs to run your website etc. The savings in India can be 20%! My auditor has told me that the tax department expects that at least 40% of your consulting revenues will be profit - meaning you can expense more than half of your income - of course, only using legal means. If you can expense out 50% of your income, you would pay tax on only the remaining 50%, which at today's rates would be 15% of your total revenues.

Let's say that as an employee you make Rs. 700,000 post tax (on Rs. 10 Lakhs) and another Rs. 200,000 in benefits. So what you need to make post tax now, is Rs. 900,000. Depending on how much you are allowed to expense that you otherwise spend anyhow (such as your web site, commute cost, internet charge etc.) you can figure out what you really need to make pre-tax: For the above figures, Rs. 10.6 lakhs if you were able to expense 50% of your revenues.

With regards to an hourly calculation : 1000 hours is perhaps ok. If you worked half of the available 50 weeks a year, you'd make 25x40=1000 hours a year.

A daily charge can be arrived at using the same kind of formula: 50 workable weeks of five days each gives you 250 days. Divide this by two for a reasonable buffer. In my example, you would charge Rs. 8,480 per day.

Multiplied the daily cost by 20 for the month's charge.

For a two-four week gig, this pricing should be fine since it's going to take you as much time to find another contract. But if you need to pitch for a contract that lasts 9 months, you might think of reducing your price for it's now more than a half years work (which is what you have considered in your pricing)

Of course, in the end, it's all about how much your customer will pay :)

The "dhanda" of Software: The Indian Perspective of the Software Business.

Thanks to Mohit, I have received my personal copy of Eric Sink's "The Business of Software". Eric is a darn good writer, a programmer, an entrepreneur and a someone who created a "Micro-ISV" for the heck of writing about it.

All of that makes him someone I want to be, except he's 38 and I want to be all of that right now when I'm 32. Go on, smirk away.

The book's an excellent collection of his writing, mostly from his blog. He talks about starting up, staying there, hiring people, paying them, negotiating with customers and in general, the transition from geek-programmer to geek-entrepreneur. You should read it if you're looking to start a business, unless you are
a) an employee of SourceGear, Eric's company or
b) someone looking to compete with SourceGear, because he paints a not-too-rosy picture of his industry.

I think I'm inspired enough by Eric to want to do something more than I'm doing; and yet, I think some of what he says needs a little bit of time to work where I am - in India. The U.S. went through a massive boom and bust in the I.T. industry; in India there's still a boom and no bust in sight. (they're never in sight, are they?)

Attrition rates are high and the job market seems to be so good that it makes little sense for the "Micro-ISV" to even think about ISVing. Or is it?

What's stopping people from going out and starting up? This is a question I asked of a group that's part of a startup network. Here's what I got:

Not yet ready.
This means someone's not emotionally invested enough to launch. Most people think this business of starting up just "happens". Something will suddenly inspire me, they think, to go out and start my own company. It's not a lack of ideas; there are still those that think selling pet-food online is a killer idea that will just have be acquired by Google. Yet, they would like to remain corporate slaves because the idea is still "in an ideation stage", which is another way to say "I'm still chicken".

Not yet ready, to me, is just an excuse. An excuse I've used too, no doubt, but that's me being a wuss. At least I admit it.

How will I survive?
A very legitimate concern for startups is: how am I going to pay the rent? Most people are willing to cut down their lifestyles - what lifestyle are you going to have anyhow, if you work 20 hours a day - but some expenses are just not avoidable. Rent, food, kids expenses and electricity payments is perhaps the bare minimum one needs.

I would plan for a six to eighteen month cash flow, depending on the type of startup. That then determines how much salary you need; if you have that as savings (and a reasonable buffer) you should be on. Otherwise, you have to add that to your funding requirements - I'll get to that later.

Don't want to quit the day job.
Let's face it: Some products need you full-time. You can't set up a retail shop by continuing to have a day job, and you can't sell software or services (in India) while working in another job. And if you're familiar with Bangalore today, you will understand that traffic is going to sap all your energy so you'll have very little time and mental energy to cope.

Most entrepreneur books and sites tell you to start up before you've quit your day job. Apart from making you live in a cucoon and postpone your launch indefinitely, what can happen is that your employer may claim it as theirs; most companies require you to sign off rights to any of your work while in their employment, regardless of whether you did it at home or otherwise.

If you got a free invite to a concert, would you bother going to it in the rain? What about if you paid Rs. 1,000? Creating a product while working full-time for someone else is usually treated like the former - I didn't sweat while making it so its loss is no big deal. Your project might just languish like the hundreds of thousands of those with the same idea; and you should actively protect against such a fate.

Meaning, if you're ever going to NOT quit your day job, set up a time frame when you will. And if you're "not ready" at that time, shut down your computer and lock it up in a garage.

Having said that, it's important to have a small cash nest to depend on in times of adversity. So working to build that nest is a good way to go; except you need to know how much you need or you'll save till you retire and then your pet-food-online-store idea is defunct because no one has any pets anymore.

Don't have the money
How can you start if you don't even have enough money to take the product forward? There are enough funding sites and blogs that talk about how to get equity funding, angel capital, seed money and the sort. I won't go into that, but will say that you can find money through Consulting and contracting: Do a few gigs here and there to get the moolah in, and phase it out as the product gets better financially.

You're going to get beat up sometime - either it'll be money, or your product features, or your customers or employees. You might as well get ready and go for it.