Log in

No account? Create an account
Previous Entry Share Next Entry
Combatting my own ignorance
One of the LinkedIn links today points to a rather pithy and direct article on the "Ignorance Anti-Pattern" for startups -- basically, saying that one of the common failings for many startups is that they don't know what they don't know, and usually don't try very hard to find out.

It's good food for thought, and I'm pausing and reflecting about it. I've actually tried fairly hard to *not* fall into this trap -- I understand the technical side of the problem fairly well (not that I have all the solutions yet, but I'm at least aware of most of the problems I'm going to face), and I've thought a lot about the business side.

Still, I'm kind of in a bubble here. The biggest problem with founding a company almost single-handed (Kate and Aaron are both substantially involved, but I'm about 75% of it so far) is that I don't have people challenging my assumptions very often, and that's an important exercise.

So here's a request to all of you. Now or later, when you see problems that I'm plausibly going to be facing with Querki -- whether they be technical, logistical or business-related -- please raise them. That won't always be comfortable for me, but it's definitely necessary. Some folks have already been doing that occasionally (my father peppered me with a fine list of questions at Christmas), and so far there hasn't been anything *too* bad that I hadn't already foreseen, but I think people are only now beginning to see the shape of where the company is going. If you see that there are competitors that I might not know about, or business traps that I may be heading for, please bring them up.

This project is, in many ways, staggeringly arrogant, and the only chance I have of success is having my eyes wide open throughout...

  • 1
Where is the post where we were coming up with specific use cases? I want one to help me do persona/general sca research.

I'd have to dig for the post, but I've been recording most of them in the wiki anyway, so it doesn't matter hugely. Tell me more about what you're thinking?

(Deleted comment)
Actually, it was thinking through the business plan that mutated this from the personal project originally intended to an actual company. But you're correct that I haven't gotten around to writing it down heretofore.

The business is almost so straightforward as to look weird -- it's a straight-up freemium play, using a mix of advertising and paid memberships.

More precisely: the business plan is designed around being honest about the modern online privacy trade-off. You either pay for privacy (and get it), or you get advertising.

I expect the majority of users to be relatively casual, with little or no data that they care passionately about keeping private. Realistically, I probably *have* to use advertising and related techniques to monetize that. That looks likely to work better than average for advertising: Spaces tend to be topically-focused, which means that advertising can be decently targeted. I don't expect anything like Google's ad rates, but it should do better than routine banner ads.

Moreover, once we move into Apps (late this year), the possibility for advertising improves markedly. Most apps have a number of natural advertisers, and a bidding-based advertising model ought to work rather well. The back-of-the-napkin plan is that advertisers would bid to place ads in child non-paid Spaces from relevant Apps.

I don't *like* the idea of making this tradeoff, and I might put it off for a while. But like I said, I expect it to be necessary.

OTOH, I despise the modern systems that pretend to be completely free, and therefore wind up compromising privacy all the time. So I'll be encouraging folks to buy paid memberships for a reasonable price. The price point is yet to be determined -- my guess is that $20/year will be the sweet spot, and I'm hoping that proves feasible, but there might be a basic $10 version as well, and there will likely be upsells for advanced features such as domain hosting.

Paid membership will gain a number of features, but the most important is private Spaces. Unpaid members will get only a limited amount of private Space (how much is TBD); paid members will get a lot. And private will mean *private* -- no reselling of the information, individually or in aggregate. Basically, you're paying for the ability to manage your ACLs properly. This fact drives a lot of the decisions about which features are member-only.

I'd far rather be using this subscription model -- in my gut, it feels more honest -- but I know from experience that only the relatively serious users will go for it. I'm hoping for 20% of users, but expecting it to be more like 5%.

This model isn't fashionable, and isn't how you build a modern $10 billion Internet powerhouse. But one of the more freeing decisions (made about 5-6 months ago) is that I'm not going to *try* to become a billionaire. Frankly, if we can get to $10 million in turnover and a decent profit, I'd probably be content, and so far there is decent reason to believe the company could hit a $100 million valuation eventually. It isn't the Facebook or Google approach, but there are scads of small companies (including LJ) that toddle along decently well using this model. So I have reason to believe it can work.

Of course, I have no idea of whether it *will* work. One of the rules for a successful startup is that you start with a plausible plan, and then adjust to reality on the ground. I'll be surprised if this business plan doesn't get significantly adjusted over the next two years. But it seems plausible on its face, and makes a good starting point.


Edited at 2013-03-06 03:49 pm (UTC)

(Deleted comment)
(Deleted comment)
The price point is yet to be determined -- my guess is that $20/year will be the sweet spot, and I'm hoping that proves feasible, but there might be a basic $10 version as well, and there will likely be upsells for advanced features such as domain hosting.

Minor aside: If you end up having more than one level of pricing, put in a high-end option that you don't expect most people to take. "Budget, Normal, Premium" is a standard choice-palette for a reason; you get a lot more people signing up for Normal.

(http://www.ted.com/talks/dan_gilbert_researches_happiness.html for one source as to why.)

"And private will mean *private* -- no reselling of the information, individually or in aggregate."

A question which it would be better to answer ahead of time than under pressure: How will you respond to subpoenas or other law enforcement requests?

(Deleted comment)
As for who the customer is and why they'd buy it, that's a hydra of an answer.

I expect the *initial* audience to be geeky early-adopters like my friends, who want to use the underlying Querki "toolkit". Querki's going to be one of the easiest ways to manage casual data, and is quickly going to wind up more powerful than most options available to routine consumers, with features ranging from limited-access online sharing to email capabilities to its loose but powerful data model coming online relatively quickly. For the programmer-geeks, it's going to be damned useful: while each of the features is individually available, some of them in combination, I don't actually know of anybody who's ever put things together properly for the educated modern consumer. There are a lot of pain points to address.

I don't expect to hit the mass market until we implement Apps, though. This is the model for taking a working Space, extracting its underlying structure, and promoting it to be shareable with others. At that point, we add an App Gallery (because this *is* a platform play, and you have to have an App Gallery), and it becomes near-trivial for non-engineers to take an App, turn it into a Space for themselves, and start using it. Basically, it simply looks like saying "I'm going to use this App", although under the hood it's quite different.

Apps become viral once we add in social network integration, which is coming relatively soon. There will be some basic table-stakes virality (such as making it easy to post when you've created a new Space based on an App); more importantly, a number of the Apps already designed are fundamentally viral, designed around encouraging users to share the data in Spaces such as Cookbooks and Restaurant Lists. If we can get to the point where we have a general, easy-to-use toolkit for building viral data-sharing Apps, that becomes *very* powerful and disruptive.

Apps gain another boost when we add profit-sharing. This is still at the back-of-the-napkin stage, but the plan is that folks who build popular Apps should be able to tap into the revenue stream from them. That provides a way for other people to build businesses on top of Querki, an essential component for any platform play. At that point, we have other people marketing the tool for us, which helps things spread.

There's more, including the plan for a potential enterprise-focused subsidiary, but that's a few years down the road, so it's too indistinct to be worth a lot of attention yet. But there's plenty to deal with in the next year, to get to the point where we're up and running as a real company.

And again: nothing is hard and fast. This is the current tentative plan -- good enough to be worth moving in this direction, but I'm under no illusions that it'll all go smoothly. This is why I'm shooting for release as soon as I can push it out the door: the functionality needs to be tested against reality early and often, and the actual details are going to depend a lot on what the users tell me. We're going to start with baby steps, and go from there...

Of course, what I *don't* have yet is a good elevator pitch for investors. Fortunately, I'm not looking for investors. I won't be surprised if I wind up needing angel money at some point, to deal with early cashflow issues, but it's quite possible that I won't. My preference would be to just build things up slightly more slowly, and operate out of revenues. (Hence, I expect to introduce paid memberships relatively fast, likely by the end of this year.)

And I'm deliberately avoiding VCs: long experience has taught me that venture money is a mixed blessing at best, and is at least as likely to destroy a good company as make it. They introduce a lot of weird problems, ranging from making the company too fat and lazy early on, and then too desperate for profits when they should be focusing on the long term. I'd prefer to just not play that game.

Hence, writing up a properly *formal* business plan just plain hasn't been a priority -- there's nobody I'm pitching it to. But that doesn't mean I've been ignoring it: I've actually spent almost as much time over the past nine months designing the business as I have the technology...

(Deleted comment)
And is it pronounced "quirk-key" or "k-where-key?"

(Deleted comment)
Have you been either (a) consulting with tpau or other UX professional, or (b) test-driving the UX on actual users who accurately represent your target customers? If not, when will you begin to do so? One of the cardinal sins I see in programmer-driven projects dealing with powerful data engines is building UIs which mimic the way they think about the data.

(Or which mimic the way the programmer thinks people should/will think about the data or accessing-the-data... only they're wrong.)

Changing UX workflow is often one of the more painful sorts of UI alterations. And it doesn't matter how powerful a data-mining tool is if people are turned off of it in the first 60 seconds of use. This cropped up at my last job, where one of the more brilliant programmers built an incredibly customizable reporting engine which exactly one person at the company ever was willing to use.

Edited at 2013-03-06 06:30 pm (UTC)

Have you been either (a) consulting with tpau or other UX professional, or (b) test-driving the UX on actual users who accurately represent your target customers? If not, when will you begin to do so?

Not yet, although I'm intimately conscious of the issue. This is why I'm driving towards alpha ASAP (was hoping for April, but will likely be May), and I'm expecting to spend essentially the middle half of the year *primarily* focused on UX.

I dearly wish I had the money to be able to simply hire a real UX expert; fortunately, I have enough experience in that arena that I can probably stumble through the process semi-adequately until I can do so. (tpau has actually recommended that I do a bit of formal UX training, to shore up what I already know, and I may pursue that soonish.)

Anyway -- yes, critical issue. This is why the alpha is going to open small, hedged with lots of caveats of, "help me figure out how to make this pleasant to use", and much of my time dedicated to working on that...

Also: Assuming Querki turns out to be a functional, useful tool: how would I ever find out about it if I didn't know you?

More relevantly: how would I ever find out about it and be interested enough to actually spend N minutes of my life poking at it, if I didn't know you?

(Where N = enough time to be convinced to sign up for it.)

Assuming Querki turns out to be a functional, useful tool: how would I ever find out about it if I didn't know you?

Initially, I'm very intentionally focused on the niches that I'm already a member of -- I have specific plans for applications to pull in folks from things like the SCA, fandom, and so on. My current guess is that I can probably get ~5000 users that way, some of whom will find the concept intriguing enough to do more than scratch the surface.

Beyond that, it's the crossing-the-chasm exercise. The too-handwavy plan is to begin turning out a lot of useful Apps, covering a variety of problems, to get the point across. Given good enough social-network integration (absolutely crucial to the project having any hope of success), that should cause gradual but steady growth, and that's really going to be the focus until early next year. By then, hopefully we'll be technically ready to handle faster growth, and will have some plans for how to make it happen.

The key is that I specifically do *not* expect most people to pick up Querki as a toolkit. The real growth is only plausible from the Apps themselves. The trick is going to be getting enough engineer-types intrigued enough that *they* start churning out Apps, to make the crucial jump from 1/week to 10/week to 100/week. Revving up that spiral is central to success, because the Apps are what will drive adoption...

  • 1