For a long time I used desktop clients such as Newsgator or FeedDemon, until their gradual demise. For me like for many other people, at some point Google Reader became the de facto cloud based, cross-device replacement. Despite most RSS hot companies from circa 2007 falling out of favor for a lack of viable mass-market business model, Google had a good shot at sustaining the format for general user with both Feedburner and Reader. Yet they actively sabotaged the former, while they just announced they would shut down the latter in July. OK, shut down stuff that’s actually useful so people can continue to ignore Google+, that makes a lot of sense. Well in the crazy world of megacorps it does. I knew years ago they would go on a portfolio cleansing purge eventually, I just didn’t expect them to shut down the good stuff.

Anyway, I “migrated” my personal selection of 500+ feeds to Feedly and I quite like it so far (technically I just migrated my reading since the feeds are still handled by Google in the background I believe). Despite a couple of error messages, the migration process was overall pretty smooth. At first glance it seems a viable alternative for the power user who needs a UI that allows high volume speed reading, for which glossy readers such as Pulse or Flipboard are ill-advised. Either the guys at Feedly had been wisely crouching for months, ready to jump in anticipation of this moment, or they’re amazingly responsive. Either way, well done. The experience on Chrome is pretty good, I have not had time to try the Android and iOS clients yet.

For the time being I plan to keep using Netvibes for work where I follow a whole different set of 900+ feeds. It’s a much more deadline-oriented pressured desktop workflow where Netvibes works OK. The Google Reader/Feedly user experience is more for catching up during downtime / off work, mostly on phone/tablet.

Update 1: header on Netvibes today: “If you’re experiencing slowdowns or feed latency, please bear with us as we work hard to handle a huge amount of new users.”

Update 2: Matt Haughey’s thoughts. I’ve been able to set up Feedly’s options to my liking in Chrome and on my Galaxy SII, to the extent that it still feels a little bit different yet as productive to use as Reader was. I haven’t tried it on the iPad yet, which I use mostly for work and to read long ebooks. The big question is whether Feedly and their “Normandy” API will prove a viable long-term replacement to store and manage subscriptions once Reader is offline. They say they added half a million users in just a couple of days. The capable way they handled Reader’s bust is rather encouraging.

Update 3: 3 months to scale Newsblur.

Update 4: Jason Shellen on Reader’s early days.

In my experience two scenarios often play out that give the customer an overall “meh, whatever” feeling no matter how great the work done by support.

1. Documentation, what documentation?

File this under “great customer support interactions that I would rather have avoided.” Most of our support requests are due to documentation that is either hard to locate, inconsistent, too abstract, or out of date. I’m thankful to the customer rep who knows what they are doing and gives me a fast, accurate answer that solves my problem, but I still had to spend more time than I’d have in the case of well-maintained documentation.

Once you become familiar with a vendor, you start to see that you didn’t run into an edge case, but rather that not only documentation was poorly written and delivered from the start, but nothing is done to tap customer support and improve documentation as a continuous improvement process either. I feel for the poor reps tasked with baling out water out of a leaky boat, and net net I’m not thrilled by the vendor. Not only have they failed to properly document their product, but they are not doing anything to even start to fix it. Big missed opportunity. Use your support to detect and address deficiencies in your documentation, so even if it’s not great, it will get better incrementally.

2. For product feedback, press 5

The second common pattern that shows customer support has been set up as a silo is support responses that go something like this: “no, we don’t do this, but this is a great idea, why don’t you go and fill in this form to send it to our developers.” Huh, I don’t work there, you do, so why don’t you? Usually there’s at play a mix of lack of internal currency given to support staff (underlings paid a fraction of what developers make), and a lack of software integration in the vendor’s internal systems (i.e. their support app can’t talk to their product management app).

Either way, this compartmentalized thinking doesn’t cut it. Submitting a proper support request already felt like work, so it’s likely the mental capital and time that I wanted to allocate to the issue is already spent and at this point I just want to move on.

This does not happen with smaller startups where developers support their own product, but as organizations grow they tend to move their programmers away from front-line support. Fair enough, but don’t make me do the same customer twice, and don’t make me feel like your product team does not listen to its own support colleagues.

Bottom line: customer support is a great source of improvement ideas for your product and its documentation. Oh, and it’s also a great lead generation channel, from disgruntled customers you can win back to people toying with a free trial that good support converts to customers. But at least most software shops probably already get that part, as it’s money directly in the bank. My point is that ultimately, integration between support and product/doc also is.

Related

• How to Handle Feature Requests – Tips and Scripts from 5 Top SaaS Vendors

Despite applying the same upstream filtering to my various social media streams (i.e. I don’t follow celebrities and other noisy crap in Twitbook+), I find their information density way too low. If anything, the current crop of self-centered, ephemeral social media show the lasting value of a good blog with a purpose.

This infodiet clean-up forced me to confront the fact that I had stopped my own blog out of sheer laziness. Part of it was technical, as my Movable Type install had gradually fallen apart and I was not eager to commit hours to try and fix a system that was obviously a dead end. And part of it was an underlying lack of writing motivation, what with my day job already implying a lot of reading, writing , editing, and thinking about these tasks. Part of it was the nagging feeling that if it’s worth writing a blog post, it’s worth doing it right, otherwise you might as well stick to 140-character one-liners.

It happens that I had to become fairly proficient with WordPress in the past year for professional reasons (aka “do more with less”), so the perspective of fixing and continuing to operate my own CMS had become less of a hurdle. (I don’t like much hosted CMSes as they feel like straight-jackets.) And there is a number of issues I’d like to write about once in a while, from international macroeconomics informed by micro “facts on the ground” (i.e. no laughable pseudoscience math-based BS), to the evolving software as a service ecosystem, to modern online newsmaking. So the motivation built up to reboot my domain name.

I don’t plan to write very frequently, but I do plan to revive a modicum of a personal presence in a space that I own. Stay tuned. Stay tuned? Who am I kidding, nobody is going to read this entry, and rebuilding even a minimal audience will take time, if it happens at all.

People who use the Amazon.com web services routinely complain about their sluggishness. In the dedicated discussion board, after someone suggested they create an XML feed to advertise the current availability and average response time, as well as planned downtime, a developer from Amazon said they’d look into it. Better late than never, I advocated something similar for Paypal in late 2001.

Sometimes I wonder whether their web services are anything more than a cool gizmo for Amazon.com. They certainly don’t give the impression they consider this a mission critical tool. Affiliates who use Amazon’s web service need to cache the data and avoid live calls that might break their own site.

10/05/04 update: Rips in the Web 2.0 fabric.

11/05/04 update: Bringing Web Services to the Masses.

02/26/06 update: Trust.Salesforce.com.

09/24/07 update: this is finally something more companies are doing now, e.g. heartbeat.skype.com.

09/01/09 update: Google Apps Status.

As far as we’re concerned we’re making sure we don’t just play around too much and actually fully use a tool before considering implementing yet another one. Right now we’re focused on Clicktale, mostly for its forms analytics. A bit slow but if you’re into analytics, you owe it to yourself to try it.

There are other tools that we’ve been using for a while, and yet others that we might test later. I’m sure I’m missing many but here’s a list:

• SEOmoz Open Site Explorer (link popularity and backlinks)
• Google Trends and Google Insights. We use this for keyword and website trends. Why two separate tools though? Plus Google Ad Planner.
• Google Webmaster Tools have been improved over the years and are now quite useful to diagnose site problems, though its interface could still be better.
• PostRank: social media content spreading. Haven’t played with it yet but I’m interested. So far we’ve used GA’s advanced features and AddThis to monitor how content propagates over social media. Meh.
• Among the Twitter analysis tools, I like Twiangulate best for its raw ambition and potential. All these tools (including Google Labs’ which gave me an error 500) are still flaky and tend to crash on non-trivial queries.
• ChartBeat. Real time stats. Would love to be in a situation where we could actually act on that. We’re not really in hot news markets anyway, but I could see how someone could waste days with this.
• For more, see the excellent 30 SEO Problems & the Tools to Solve Them (part 2), part 1

Google should really open up GA in a way that lets it hosts third-party plugins. I want link popularity analysis or social media or forms conversion or split tests all in one place! Tools that compete for attention don’t scale. Why isn’t Google Trends integrated in Analytics? Why do we need to manage credentials and dashboards for so many apps?

The web application world, especially in the thriving marketing space, needs its Microsoft Office moment. It’s the suite, stupid. But it comes with a twist: let me build up the best-of-breed online suite that I want, that I’m actually going to use, and that I can afford.

Update: GA App Gallery definitely along the lines of what I was talking about. The have WordStream, among others.

Despite the buzz, I’m not really excited about Google working on a browser or IM client, though I definitely can imagine them buying Trillian and giving its Pro version for free. I guess they’ll negotiate interop with AIM and they might even force the hand of Microsoft and Yahoo, which would be a Good Thing.

But looking at Google Desktop and its local web server comes a more intriguing thought. How about partnering with or acquiring a large ISP/WISP (say, Earthlink) to deliver an affordable service bundle with symmetrical bandwidth, static IPs, reliable DNS, and self-publishing with Blogger, Picasa and Hello. Let millions of personal web servers bloom and piggy back on that big wave of user-generated content.

Google would basically re-index their customers’ sites (just a directory on their desktop really) on the fly, and share the results with the rest of the world (or not) based on user settings (do not confuse the wedding pictures and the honeymoon sex tape, ok). And now it makes sense to give software for free because you have other ways to bill consumers and learn about them. How’s that for increasing targeted ad inventory while diversifying your revenue sources, and wiring yourself into people’s life as well as within the fabric of the internet?

There might also be a side business out of caching in there for Google, they already have part of the infrastructure and would only need to move to the edge. (Hmm, ok, looking at Akamai’s $150M in yearly revenue, caching is not that exciting from Google’s perspective at this point.) Add P2P (the Hello angle) to that to take care of content propagation so that publishing something popular is not asking for a DDOS.
Anyway, imagine the landscape 5/10 years from now with ubiquitous PDA/cam/phones/whatever, lots of connectivity all around, more occasions and ways to generate content and to put it online instantly. Uploading pictures to a damn server with restricted storage just to share them with friends and family is akin to going to the telegraph office to send messages. It’s just a transient state in the infrastructure that doesn’t make any sense in the long run (provided we eventually get decent security on the desktop).

Google is not going to win against Microsoft or even decisively beat Yahoo by going through predictable motions. GBrowser is just a way to wave a red flag at Microsoft with “please come and squash me” written on it. I find it funny that the same people who get all wet about the GoogleOS are claiming that operating systems are a commodity and there’s no money in there anymore. So why should Google do it then?

Instead, Google has to keep being disruptive and unpredictable. Microsoft never made much out of its broadband investments, Yahoo is humming a boring song with SBC, AOL TW is toast. Incidentally all seem to confuse the Internet with TV. Maybe Google could really turn the tables. It’s all about empowering end users…
Right now even Altavista is better than Google at indexing media content. A first telling sign would be a Google Images that doesn’t suck. We’ll see (literally).

In other news, I’m amused to see CNet slowly wake up to search engine optimization techniques that we small fish have been using for years. First they started stuffing keywords into their URLs, now they look at their referrers to welcome Google users and suggest other relevant stories besides the one displayed. Wow, leading edge!

01/19/05 update: CNet: Google wants ‘dark fiber’.

02/07/05 update: Google Now a Registrar. Continuing with my theory, Google will give away free domain names as part of their Broadband offering. Much more elegant and less arcane than dynamic DNS.

05/05/05 update: Google Web Accelerator.

08/18/05 update: Free Wi-Fi? Get Ready for GoogleNet (speculation too).

09/09/05 update: Cerf’s Up for Google.

11/25/05 update: Google’s Shipping Container.

02/15/06 update: All Your Information Are Belong to Google.

09/24/07 update: Google plans new undersea “Unity” cable across Pacific.

01/24/08 update: Broadband Capacity Is The Alternative Minimum Tax Of The Web.

02/10/10 update: GOOG’s experimental fiber network

02/17/10 update: How much should Google be subsidized?

03/17/10 update: Google traffic

Realizing your family, yourself, and your property don’t have a scratch: priceless.
My thoughts to the families of the deceased, wounded and homeless. Te quiero Chile.

Update three years later: I had posted this after the 8.8 earthquake on Feb. 27, 2010 in Chile. We live about 350 km (215 mile) north of the epicenter, so we felt a pretty long ~M7 earthquake on the Richter scale. Not tsunami traumatic, but definitely scary nonetheless (the noise!). This will remain a strong memory for life for sure. Oh, and the aftershocks (we’re in the upper part of that collection of 458 aftershocks, with a bunch of M5s-M6s right where we live).

On the web publishing tech front, we’ve been working with ExpressionEngine for a bit more than a year. We like it for the most part, but still have several unresolved issues:

• 404 pages that actually work in all cases. Documented issue with some leads into a resolution, on our to-do list.
• REST search results with clean URLs. The default EE behavior is a major design flaw IMO, but don’t all CMSs handle search terribly in their own ugly way? Solspace recently released a beta module that looks promising, we have yet to play with it.
• Making dev/staging/production management and deployment easier. We use SVN (with Unfuddle) and are looking into moving to Git. Some people have posted how they do it but it seems no methodology is entirely seamless because EE insists on storing some settings in the database, which makes portability across servers more of a chore.
• We have struggled with comments/spam but contrarily to the previous points, it’s not necessarily that there’s a problem with the product, we just under-invested time to get it right. Obviously still on the to-do list.

Still, EE is way more manageable for us than our old model of using one WordPress installation per site, thanks to EE’s multi site module. I guess I wish more advanced features came right out of the box. We still have a lot of work ahead to execute all our ideas. That sort of stuff just never ends. Just like fixing and improving a house I guess! (Half horrified, half amused exclamation mark.)

Nokia SIP call -> Claro 3.5G network -> internet cloud to the US -> Vonage -> more internet cloud back to Chile -> Tutopia DSL -> Cisco ATA 186 -> Philips VOIP841 wireless phone (for whom Vonage is my “landline”, I also have Skype running on it).

There’s probably 20,000 miles worth of roundtrip involved, with data packets going through air, copper and fiber. Try to visualize it.

My wife who was upstairs picked up the phone, and my unscientific measure of the lag involved was maybe 0.5s. I admit my whole internet/VOIP setup is a bit complex: cable/dsl/3G, IP clients include laptops, a smartphone, a NAS (downloading 24/7 in the background), three VOIP providers, an Xbox360 (well it’s toast right now), two internet mp3 players… But the bottom line is, it’s un-freaking-believable what we can do with IP devices these days.
When I was working at AT&T in 1995, the company had very visionary products, from what would become wifi, to WAN video conferencing to early tablet PCs. None of it worked too well, and the party line was that VOIP wouldn’t work through IP over Ethernet: “We tried it, no can’t do. Dude, you want voice/video over even just your LAN? Go ATM.” Well, 14 years later, there’s maybe 10 miles of ATM involved (it’s used by DSL providers) but the whole thing is IP traffic.
Update: Wow, a well-written Wikipedia entry? There’s hope after all.

I haven’t been able to get my new 3G modem to work when plugged into my Draytek 2910G router yet, though their supposed to be compatible. Gotta get it to work as a backup in the rare case dsl and cable might be down at the same time. (Me, paranoid and obsessive compulsive? You bet.) I’d also like to get Teliax to work through my other router (Draytek 2600VG), no dice either. More nerdrage needed to get everything fully working I guess. (SIP settings and registration management on Nokia phones are a bit screwed up, so there’s been some bitching involved to get where I’m at right now.)
Update: I’ve updated my Grand Central account to Google Voice. I can route my Google Voice to either/or Vonage and Teliax (also meaning, the devices of my choice, at my desk or on the road), get voicemail to text or recorded calls for free. Amazing.