Venture Capital Blogs OPML file · 700 days ago by Ken King
Andrew Fife posted a list of Venture Capital blogs that’s been getting a fair amount of attention lately. I was already subscribed to a bunch of the feeds, but found some interesting new sites on the list.
However, it took a long time to go through them all especially since the way the RSS links interacted with Firefox was all over the map. I finally ploughed through it this morning and got the information transferred into NetNewsWire.
To save other people some time and effort , here’s the VC blog OPML file I exported from NetNewsWire. Right-click and save the file, then use the import feature in your newsreader to add all of the VC blogs at once.
I’ve also uploaded it to Share Your OPML.
In another interesting development related to this list, Brad Feld used a new FeedBurner feature to create the Venture Capital Network – a spliced RSS feed that combines the output of the 13 VC bloggers he’s invited to the network thus far. [via Fred Wilson]
Comment [1]
Enablr buzz: BusinessWeek, Inside Digital Media, Rogers Wireless · 794 days ago by AJ Kandy
BusinessWeek Online’s Tech Beat blog covers Transcribr in a piece by Olga Kharif.
Enablr honcho Ken King was interviewed by former securities analyst-turned-new media guru Phil Leigh of Inside Digital Media for their weekly streaming WebEx presentation, available as an audio stream here.
Or read the Transcribr transcript here.
Some months ago, Ken was interviewed by Rogers Wireless about podcasting (along with Montreal art gallery owner and podcaster, Chris Hand). The interview appeared this month as a mailing sent out to Rogers Wireless subscribers, in conjunction with the launch of their new podcast-to-cellphone service.
Click the image to view a scan of the article in full.
Bibliography Generator · 890 days ago by Ken King
I’m a little slow on the uptake, but here’s a link to a simple little bibliography web-service created by Jonathan Otto in response to a Seth Godin post about bibliographies being stuck in the past
Jonathan admits that the service doesn’t solve all of Seth’s issues, but it’s still a pretty damned cool thing, and reinforces my recurring point about taking action. When you’re exposed to an idea, and have the time/knowledge to do something about it, DO IT.
Enter a single ISBN, or a series of them separated by commas, and you get back bibliography citations formatted according to one of four standards: MLA, APA, AMA, or Chicago.
Mechanical Turk: what's it all about? · 945 days ago by Ken King
I first became aware of Amazon’s Mechanical Turk service when it was mentioned in this article about peer production on Union Square Ventures’ blog. We thought it might be an interesting way to source labour for our clients’ transcription needs, so I set out to experiment with it a couple of weeks ago.
I looked around for a bit, but couldn’t find any HITs (Human Intelligence Tasks, their term for things that humans can do better than computers) other than those posted for A9, which consist of photo recognition tasks. Essentially, you are shown five pictures and asked to identify which one best represents the business name or street address listed at the top of the page. The promised payment for successful completion of this task is 3¢ U.S.
I put in about 10 minutes my first time out, and got mixed results. A large proportion (almost half) of my submitted tasks were rejected, and I spent a lot of time trying to scrutinize some pretty inscrutable pictures. I figured out some simple rules to get more efficient about it, though, and took them into my next work session:
- Preview the task
Before accepting a HIT, look over the images to see if the “right” answer will be easy to discern. If not, skip to the next HIT. - Focus on retail businesses
It’s a lot easier to recognize and identify a photo showing a retailer’s storefront than it is to look for a street address for a dentist’s office in suite 123 at 4536 Main Street. The pictures on A9 were shot with a video camera out the side of a moving vehicle, so reading street addresses is a pain in the ass and a waste of time. - Skip HITs with multiple “right” answers
The algorithm for determining the correct answer is a majority rules process based on multiple users’ responses. Working in that framework, you’re not just trying to find the right answer, you’re trying to game the system by figuring out which answer most people would choose. You’re penalized for “wrong” answers both by having wasted your time answering them, and in your qualification rating which might prevent you from getting access to more lucrative gigs down the road.
The results
With those thoughts in mind, I committed a solid half-hour to working at these image recognition tasks as fast as I could. Over that time, I managed to submit 25 HITs, of which 7 ended up being rejected despite my best efforts to avoid any ambiguous photos. Even if I managed to eliminate the rejected submissions, 50 HITs/hour = $1.50/hour.
At that rate, I can’t be bothered working on a larger sample size – has anyone else got any experience to share?
UPDATE
Okay, apparently I suck. ;-)
The users of this mturk board are getting to as many as 800 HITs in an hour, with reasonably small rejection rates. I bow to their superior skills and tools (there’s much talk in the forum about a script that speeds up the processing), but even at that blistering pace the pay is $24/hr.
DIY transcription - the costs · 945 days ago by Ken King
For many, hiring us to transcribe their podcast isn’t a viable solution. I said this in response to an earlier comment:
... without the potential for profit, any expense is difficult to justify – profit puts the R into ROI. Our service is primarily targeted to podcasters who can see a return on their investment, generally those using podcasting as a promotional vehicle for their businesses.
This extends to podcasters who plan to earn money through Google AdSense and other advertising vehicles that are dependent on context.
For those who don’t fit in the categories above, doing it themselves could be an attractive alternative. However, it’s important to think through the costs before going that route:
A professional with appropriate specialized equipment and software averages about 3-4 minutes to transcribe one minute of audio. I have done transcriptions on my own (interviews performed for KMA+C clients), and so have several members of my staff. We averaged closer to 10-15 minutes per minute of audio. For us, the opportunity cost of doing it ourselves was higher than the cost of outsourcing the labour, and so we went with a transcription service.
Each person or organization needs to do its own assessment of the opportunity costs and decide whether it makes sense to do it yourself. For some, unfortunately, the decision may even be not to do it at all – the incremental value created through transcription may not be enough to justify any expense or effort.
Cost model for transcription · 945 days ago by Ken King
I recently posted a response to Michael Arrington’s thoughts on podcast transcription.
In that post, I discussed some background about our pricing, which focused primarily on the topline and questioning whether it would make sense to provide transcription services for 33¢/minute as per Michael’s proposal. I’d like to elaborate a little more on the cost model for transcription and the reasons why I believe $1/minute is currently the market rate.
At first glance, $1/minute seems high – “Wow, that’s $60/hour!” However, there’s a lot more to the story, and I’ve attempted to tell some of that story below:
Audio time vs Transcription time
The first thing to consider is that the rate is based on audio timecount, not the actual work required to transcribe the podcast. A professional transcriber with appropriate equipment and software typically transcribes one minute of audio in about 3-4 minutes. So, right off the bat the topline is $15/labour hour (25¢/minute) in the best case scenario.
Labour costs – contractor model
In my experience, most service businesses have client billing rates at least 3-5 times their labour costs for contract labour that is only paid for billable hours, and therefore purely a variable cost. This is simply a rule of thumb, but the differential exists to ensure that the company has enough gross margin to cover marketing, R&D and other overhead and, of course, profit for the owners.
Applying that model here, we’d end up with a contract rate of at most $5/hour. Bear in mind that this is for freelance work with no guaranteed paycheque and requires investment in specialized equipment and software in addition to having a computer and broadband access. Freelance rates typically range from 2-4 times fulltime wages (which helps compensate for the lack of guaranteed hours). We’d therefore be looking to do the work in a country where skilled, educated labour appropriate to this work would normally earn $1.25-$2.50/hour at a full-time job.
Labour costs – full-time staff model
The alternative method is to work with full-time staff, lowering variable costs but increasing overhead. If you start with one full-time person doing transcription, the cost would work out as follows:
Start with 2000 hours available (50 weeks x 40 hours/week). Many service businesses use 80% as a target for billable hours, but the real number is closer to 50%. Using 70% as a middle ground yields 1400 billable hours per year, which translates to 21,000 audio minutes (4 labour minutes per audio minute) or $21,000 in gross revenue at $1/minute.
Evaluating the trade-offs
If we were to go with an assumption of $2/hour as full-time labour cost, the cost of transcribing one minute of audio would be 13¢ ($2/hr=$3.3¢/minute x 4 minutes of labour) vs 33¢ for contract labour ($5/hr = 8.3¢/minute x 4 minutes of labour). The break-even point for going with full-time staff instead of contract labour would be 20,000 minutes of audio (20¢ differential in gross margin divided by the full-time salary of $4000/yr).
Considering how close this is to the employee’s capacity (21,000 minutes of audio/year), it’s difficult to make the case for full-time staff in this situation, especially given other considerations such as being completely dependent on a single source for the work.
Sources of Labour
This raises the question, then, of where the labour can be sourced at rates that make $1/minute viable (let alone 33¢/minute). Dr. John Rutledge posted some information about manufacturing labour rates. Bear in mind that the $1/hr and $2/hr rates listed for India and China respectively are for manufacturing labour. I think it’s fair to assume that the rate for skilled labour would be somewhat higher, and so we’d be pretty close to the bone in paying the rates discussed above.
Transaction costs
Payment processors (like PayPal) want their cut of the action. This typically works out to about 4-6% of the transaction. For example: the starting rate for a low-volume business at PayPal is 30¢ base fee, plus 2.9% of the transaction. For a $10 transaction, that works out to 59¢, or 5.9%.
Conclusions
There are none, really, other than to provide some backup for the existing market rates for transcription services.
Sample transcript · 945 days ago by Ken King
A few people have asked about seeing a sample transcript. Here you go.
The client (Alex Barnett) also posted it to his blog in a slightly different format and plans to review and edit it at a later date.
Go Flock Yourself - points well taken · 945 days ago by Ken King
I previously pointed to the Go Flock Yourself review of Transcribr and Enablr for its humour, but there were also some good points in the article and comments that deserve more thought.
The missing e
We know it’s goofy, have said so since day one.
Perpetual beta
Christian Montoya’s comment about our beta tag is funny and on the mark – there are far too many sites with beta splashed all over them. I’m torn about it though, because I happen to believe that there’s always something left to be done, whether it’s with software or any other good or service. Keeping the tag there is a reminder that we should keep looking for ways to improve. There’s a good chance the beta will disappear at some point in the near future
Why don’t you just type it in the first place?
Another comment (by Stu) suggests that “Sure, you could get someone to transcribe your podcast; alternatively, you could stop twatting about with a microphone and just type it in the first place. You know, for free.” I agree – Transcribr was, in part, a response to my frustration at finding more and more content locked up in an audio format that isn’t convenient for me.
However, economists will tell you that there’s a cost to everything, nothing is “for free”: the reason many people podcast rather than write is that it’s easier (costs less) for them to speak than to type things out. For some, it also provides higher value, because it allows them to convey emphasis and emotion without kludgy formatting, acronyms and emoticons: if someone is truly ROTFL, you’d hear that in their podcast and newbies wouldn’t have to trundle off to wikipedia to find out what the acronym means.
Accessibility
Also in Stu’s comment, the fact that nobody seems to give a shit about the fact that some people are deaf: Almost a year ago Andy Carvin posted this piece about podcasting and accessibility in which Steve Gillmor admits that he hadn’t thought about accessibility. I don’t think that anything has changed much over the last year, but I also think that most podcasters would agree that accessibility is a good thing.
The question then becomes whether they value it enough to invest time or money into the solution. For many, the answer may be no. However, improved accessibility could be the unintended consequence of podcasters whose motivation is primarily to improve searchability and ease of use.
Don’t believe the hype
Not limited to this article, the entire Go Flock Yourself site is pretty hard on the Web 2.0 hype, and rightly so. I witnessed the last bubble close up and see the hype for what it is. The best comment I’ve seen about Web 2.0, though, is from Paul Graham. The money quote for me is:
Web 2.0 means using the web the way it’s meant to be used. The “trends” we’re seeing now are simply the inherent nature of the web emerging from under the broken models that got imposed on it during the Bubble.
He also says that we’re not in another bubble for rather prosaic reasons:
The reason this won’t turn into a second Bubble is that the IPO market is gone. Venture investors are driven by exit strategies. The reason they were funding all those laughable startups during the late 90s was that they hoped to sell them to gullible retail investors; they hoped to be laughing all the way to the bank. Now that route is closed.
I agree with that assessment to a point, but I think what he’s missing is that many of the businesses being labelled as “Web 2.0” don’t need VC investment and will quietly disappear when they realize there’s “no there there” rather than imploding noisily the way all the bogus dot.coms did. Perhaps that’s not a bubble in the strictest definition of the term, but it’s close enough.
We flatter ourselves in believing that Enablr is not one of the businesses that will disappear because we’re not trying to feed off that hype: we’re just focused on providing pretty simple, straightforward services, on making it easy for people to use, and on making it profitable.
Selective hearing - more feedback · 945 days ago by Ken King
There’s a very funny review of Enablr and Transcribr over at Go Flock Yourself
Ok, this service it might not be totally without merit, but the simple fact that omit the last ‘e’ in no less than 3 words really irks me.
Lets grade 2.0 these guys, shall we?
With my selective hearing turned on, I’m going to take the possibility that the service might not be totally without merit as a compliment and say thanks for the shout-out.
You should, however, check out their 2.0 grading sheet, in which points are awarded for (among other things) leaving the e out of words, beta, and “big, stupid fonts”. We missed a few points by not making the RSS feed for this blog more obvious, we’ll have to try harder. I looked for other examples of this grading sheet, but it appears we were the first. We were also awarded a Web 1.98 button
Seriously, though, the article and comments raise a few things that deserve some thought – I’ll post that separately.
PS – also found through Go Flock Yourself, a Web 2.0 bingo card
DIY transcription · 953 days ago by Ken King
For many, DIY transcription could be the way to go. O’Reilly has a how-to article on turning your mac into an audio transcriber and includes a bundle of scripts that emulate the tools used for professional transcription.
As a bonus, the author threw in a hack to get rid of the QuickTime Pro nag screen.