:Ben Metcalfe Blog

Archive
October, 2009 Monthly archive

Yahoo! have been mopping up the PR fallout from the stripper show that occurred during the Yahoo! Taiwan 2009 Hack Day (if you missed the story, check out Violet Blue or Simon Willion’s posts).

Well, the plot thickens as the most damning photos of the event have mysteriously been removed from public viewing. Flickr user CocoChou had taken probably the most damning photos of the the stripping event and uploaded them to his Flickr set of the Hack Day under a Creative Commons license.

However, the 4 photos – which many blogs were embedding and linking to – have been made private in the past few hours, although curiously all of the other 72 photos from the event are still public.

It’s not clear whether Yahoo!, which of course owns Flickr, has put pressure on CocoChou to remove them or whether he removed them independently of any interference. However, it’s difficult to think of reasonable motivations why someone would remove these 4 particular photos from their original 80 and keep the rest up. I have, of course, emailed CocoChou to find out – and I’ll update this post if/when I hear back.

Fortunately I downloaded copies of these photos before they were removed, and in the public-spirited nature you’ve all come to know and love, I’ve re-uploaded them to my Flickr account. Of course I am able to do so legally as they were originally released under a Creative Commons license, which I have perpetuated:

Lap dance from Yahoo! Taiwan Hack Day 2009

Lap dance from Yahoo! Taiwan Hack Day 2009

Lap dance from Yahoo! Taiwan Hack Day 2009

Lap dance from Yahoo! Taiwan Hack Day 2009

Read More

Metric-orientated user acquisition is definitely the hot topic of the moment here in the Valley (along with “frictionless customer conversion” as my rad friend Ethan Bloch of Flowtown would say)

Exit door

But as we optimize funneleing and conversion for user acquisition within our startups, how many of us have a solid user relinquish process for those users who might want to leave?

As a technologist and generally curious geek, I like to check out most new start-ups that are launched. In fact it’s kinda my job to, and to that end I probably create at least two or three new accounts somewhere a day.

I visit, I sign up, I create an account (“just username, password and email address!”). But that’s the beginning of a probably well-oiled slippery slope. My interest is piqued, I upload my photo, fill out the profile data, connect my Facebook, etc, etc…

But what if I’m now done? “Nah, not interest in what you are doing”. Or even more importantly “Er, um!.. I’m put off by this weird thing you’re doing in your site. I don’t like it so I’m outta here”.

How many sites actually let me delete my account and leave?

Or conversely, how many sites have stubs of my personal data sitting on their servers because there is no easy way for me to remove it – despite being clear I no longer wish to use their service?

Now, I don’t consider myself some crazy privacy whack. I just think it’s pretty reasonable to suggest that if I really have absolutely no interesting in using your service at all then I would like to know that you will completely remove my data and forget about me. Data Portability ‘Removability’, if you will :P

Two examples from the real-world…

Earlier today I decided to delete my Tangler account (nothing personal to Tangler, I just don’t use it and it sends me a digest email every week that is just spam to me). I logged in but found no “delete this account”, so I twittered my frustration. Later on Rai from Tangler @replied to me to say that this could only be done via email.

That seems like a FAIL to me.

On the other hand I had a similar-but-positive-outcomed experience with Dropbox last week. Having unshared all of my computers from my account I still had 2gig of orphaned personal files in my storage account that were proving difficult to remove. Bug or user-error I wasn’t sure, but I decided I just wanted to nix my account and start again.

To DropBox’s credit had a “delete this account” option, accompanied by very clear warnings that it was an irreversible decision. They even had a data-capture form to give me the option of explaining my reasons for leaving/deleting my account – which someone personally followed up with me when I mentioned I was having file deletion issues.

Top marks on responsibility, implementation best practice and most optimized reason-for-leaving collection mechanism (alluded to in #5 in this great blog post on customer feedback)

And that’s all this really as to be: a “delete my account” button at the bottom of your settings/account profile page, a confirmation box and perhaps some way for the user to explain why they want out. On the back-end, a quick purge of that user’s record and perhaps a separate archived audit log so that if a backup is restored deleted accounts can be consolidated.

A user relinquish strategy is good for your business

Good user relinquish practice is not only the fair thing to do for your users but it makes sense for business.

What value is there in holding all of this information about users that no longer wish to use your service? Depending on the nature of the service you may even be provisioning resource for these ghost users – resource that you will never see a return on. And VC’s/boards don’t want to see exaggerated raw account numbers, they (should) want to know monthly uniques, return visitors, etc.

And if we get into a %age game, removing users from the database who have totally left the service will actually increase the %age of your userbase that returned in the last month! :P

There may also be boring data-storage compliance issues, especially if you trade physically in Europe.

So, that leaves the question: what is your user relinquish strategy?

[photo CC Image Zen]

Read More

Google’s GMail blog has some “handy” advice on how pick a good password to project your email account.

Don’t use dictionary words, use mixed case, your eldest kid’s name is a bad choice, etc etc. Yeah that’s great.

But the much bigger security issue I fear is that my GMail username & password is also the same username & password for:

  • My calendar (Google Calendar)
  • My confidential documents (Google Docs)
  • My credit card (Google Checkout)
  • My website’s analytics (Google Analytics)
  • My RSS feed admin (Feedburner)
  • My phone number, voicemail, IM’s (Google Voice + GTalk)
  • Some experimental projects (App Engine)
  • My photos and videos (Picassa and YouTube)
  • + more (see your list of Google services you use)

Given the legitimate places you need to put your username and password in order to access your email (ie your email client, which might be sending it in the clear each time it fetches mail), is it too much to rely on it’s security and integrity for all these other ancillary Google Services?

I am a strong believer that you shouldn’t give your Google username and password to ANYONE for this reason. It pains me to have to give it to RIM but it’s the only way they can push email to my Blackberry.

Security through segregation

It’s really about time Google separated GMail, and perhaps GTalk, authentication from the rest of their properties. At the very least I’d like to see the ability to create a separate password for IMAP/POP access that I can enter into my email client and give to RIM that doesn’t give access to the rest of my Google Account.

However, as Google becomes an ever more vital and relied-upon part of our online workflow (see how many services I use, above), I wonder whether there would be value in offering an optional RSA-style keyfob to help protect access – perhaps for a $20-$50/year fee. I know I would pay, and that PayPal have been offering a product like this for some time at $5 a fob.

Read More

I don’t subscribe to the “everything muse be free” meme that basically ignores the intrinsic value a product or service gives you. If a product or service provides me with a real value then I am happy to pay for it – either through purchase/subscription or from being monetized via ads/usage data etc.

But I’m surprised at just how expensive some of the darlings of the Web2.0 SaaS era work out to be when used at scale.

Like a crack dealer, giving you the first hit free, most of them offer a “free” plan that is clearly designed to be severely limited the moment things begin to work out for you and your business takes off. There’s nothing new with this way of doing business, but have you seen just how much your hits costs once you get addicted?

Two examples that are particularly of mind are Freshbooks and Harvest. Both are great products; built by great people I have had the honor of meeting over the years.

Time tracking service Harvest starts out at $12/month ($144/year) for a single user but at Swordfish Corp there are now three of us, requiring the 5 user plan @ $40/month ($480/year). Not much change short of $500 seems pretty expensive for a year of time tracking.

Invoicing service Freshbooks has a free and slightly limited option for individuals but a company of three would need to use the 3-staff plan @ $39/month ($468/year) but I notice that once we take on a fourth person we would need to skip to the 10-staff plan @ a jaw-dropping $89/month ($1068/year).

When researching these plans, I’m also considering what my future business needs are. With services like these, I want to pick providers who can scale with me as my business (hopefully) grows.

I should point out that one way of getting around this is to share accounts, but for time tracking this doesn’t work and for invoicing, everyone at Swordfish does their own invoicing on their client accounts.

Now, I’m not against paying for these kinds of services in general. Between myself (personally) and Swordfish, I have paid subscriptions to NolaPro (Hosted accounts package), Shoeboxed (receipt and business card data entry) and Flickr.

And I’m not saying that it’s not worth $480 a year to the company for good time tracking. I’m just saying I’m not sure a service like Harvest is offering me $480 of value a year over and above using a simple Google Spreadsheet created in 20 minutes, for free, and shared within the company.

I’m a fan of the Freemium model, but if it’s going to work the numbers can’t exponentially increase as your usage increases – it’s not fair (a form of bait-&-switch from the free accounts) and it’s also not reflective of the true cost of SasS where the cost should exponentially flatten out at scale.

Read More