DeviceReady: Test Your Android App Across 25+ Devices

I keep hearing from lots of folks that Android fragmentation is a big problem. When developing apps for Android, you have to consider the various devices, versions of operating system, form factors, etc. Sounds like a nightmare.

I've been thinking lately about one possible partial solution:

A service that enables Android app developers to submit their app and get screenshots of how the app looks and performs across the most popular Android devices.  Not an emulator.  Your app running on each of those devices.

Like Selenium + Browsershots for Android.

To test the concept, I quickly came up with a name (DeviceReady), threw together a landing page using Unbounce, shared the idea on Hacker News, and purchased about $75 worth of Google Adwords.  The results after 3 days:

Pretty good, right?

I think this is a winning idea because:

  • A company serious about their android app really should test it across dozens of devices.
  • Presumably they are spending a lot of money building and testing their app.
  • The cost of the service, at, say $100 per month (5 submissions) or $25 per submission would be negligible.
  • It is unreasonable to think that a small or mid size android dev shop (most of them) would buy dozens of handsets and manually test on each.

Idea: Screenshot Archive of Home and Landing Pages

Learning about MOAT ad search today (via TechCrunch), I was reminded of a company called Who's Mailing What! that I stumbled into while doing some Zumbox R&D.

Who's Mailing What! is a really simple and useful low-tech idea: scan the direct mail (paper) that's sent to folks around the country.  Make this database browseable and searchable to anyone who might want to see, say, exactly how Geico markets its products to consumers.  Really useful to other direct marketers.

Why not offer something similar for web pages?

Take screenshots, at regular intervals, of home pages and landing pages of leading web companies.

Not just archive.org (too coarse: Spreedly vs. Recurly vs. Chargify)*.  I don't care about links or preserving functionality, I just care about layout, wording, button design, calls-to-action, etc, which have presumably been refined over time by these companies.  Most modern web companies respect that home page and landing page design is "a process of ongoing improvement," and employ analytics and A/B testing tools (like unbounce, which is awesome BTW) to track and increase conversion rates.

I suspect that enlightened web marketers (folks who design home pages, landing pages, manage search engine marketing campaigns, display ad campaigns, e-mail marketing campaigns, etc) would pay for a subscription service that showed them how (and when!) their competitors designs have evolved (again, changes to layout, wording, buttons, calls-to-action, etc.).

* P.S. Any way to link to multiple web pages?  One link opens n pages?  Like how Hipmunk enables you to open multiple tabs on their site?


RocketMenu: Unobtrusive Search and Reference Tool

This is the second post in a series where I revisit an old idea that is "sitting on the shelf." This idea is from November 2005.

Remember 1-Click Answers (Formerly "Gurunet" and "Atomica")?  Babylon Translator? These tools (from back in 2005) were great because they provided useful reference information on-the-fly. For instance, with 1-Click Answers, you could highlight a word in any application (on the web, within Microsoft Word, etc.), and get a definition, synonyms, antonyms, etc.

With Babylon, you could highlight a word and get an instant translation in a number of languages.  Babylon was especially cool (from a technical point of view, anyway) because it used OCR-like technology and even worked with images (not just plain text).  That is, you could get an instant translation of a word that was inside an image.  Pretty nifty.

I came up with RocketMenu - a more flexible search and reference tool:

  1. Users select any word or phrase with their mouse.
  2. A small icon appears beside their selection.
  3. Moving the mouse over the selection displays search, dictionary, thesaurus, and other results customizable by the end-user. Unobtrusive text-based advertisements appear beside the results.
  4. Should the end-user click on one of these advertisements, both RocketMenu and the Site Operator (from where the original word or phrase was selected) receive a small portion of the advertising revenue.

Here's an early design:

Here, I just stuck Google Adsense to the right of the menu.

Later, I improved upon the model by removing most of the advertising and instead relying instead on affiliate revenue (that is, many of the links, when clicked, would generate revenue to RocketMenu).  I also made the menu items context-aware.  Items would become bold if the search term matches a pre-defined context.

In the following example, the search term is "www.selectsense.com" and the menu items "WhoIs", "Web Archive," etc. have all been set to become bold when the input is a web address (ends in ".com" or ".net", etc.)

RocketMenu was designed to be a desktop app (like 1-Click Answers and Babylon). Additionally, web site operators could "embed" RocketMenu into their web sites, and, in doing so, we would share our revenues with them.

 

The idea is that every word or phrase on the computer screen becomes a vector for:

  • more information (for users)
  • additional revenue (for web site operators)

I envisioned a home page with a very direct call-to-action ("Download RocketMenu Now").

My pitch to web publishers was centered around easily earning additional revenue.  Further, I was trying to run with the idea that web publishers were inspiring lots of revenue-generating searches on Google, Yahoo, and other leading search engines.

Meaning, you read about something on some web site, then you jump over to Google and search for something related to what you just read.  Chances are, Google monetized that search you just did.  The web site that inspired your search earned nothing.

 

Here is a more complete set of wireframes, if you're interested. You'll see that I was using a mix of Visio and pencil & paper for my wireframes and mockups at the time (now I use Balsamiq Mockups almost exclusively).

RocketMenu Version 1 Wireframes

Over time, a number of companies have emerged with similar products.  Among them are Vibrant Media, HyperwordsApture, Wordclick, and I'm sure many others (I haven't taken a look in many years).  UPDATE: Check out Highlighter - they have a really interesting approach.

This is one of those projects that I abandoned simply because I got busy with other work.  There might still be something here, although I am sure many details would have to change.  Maybe a web browser extension or plug-in? (see Polaris Insights Chrome Extension for a utility that displays info about [mostly startup] web sites).  Maybe a tool that is completely context-sensitive (that is, only relevant menu items are displayed)?  Maybe something like this for mobile devices (iPhone, iPad, Android)?

I'd love to hear any ideas that you might have.


Firestorm: Preserving and Sharing Old Photographs

This is the first post in a series where I revisit an old idea that is "sitting on the shelf." This idea is from October 2003.

It was October 2003, and Southern California was burning. I was sitting at home with my soon-to-be mother-in-law and we were watching the news. All of the news stations were covering the situation with panic: "Firestorm 2003"!


(Note: Thankfully this isn't me. Just a representative video I found on YouTube)

The news teams interviewed a number of folks whose home had been lost. They were devastated:

"What am I going to do!? I lost everything. My photos. My memories. "

Photos. Photos. Over and over again, in every interview. While I completely and totally understood their situation, I thought to myself: this is ridiculous. People need to stop worrying about losing their photos.

Here are my first thoughts about a solution (from October 2003):

If you only have a few moments before your home is engulfed in flames, what possessions will you spare from destruction?

Family Photos? Videos? Other physical keepsakes? This is what everybody says! In this day and age, I find this response almost ridiculous!

These possessions are often considered "priceless." How much would the average upper-middle-class family pay as an insurance policy against the destruction of these possessions? $5,000? $10,000?

It could have some very interesting business potential.

These items could be processed in a lab that would:

  • archive the images as digital images (say, with scanners)
  • archive drawings / paintings / physical objects as digital images (say, with cameras)
  • archive the video as digital videos (with analog to digital duplication equipment)
  • store the data in natural-disaster and fault-tolerant data centers.
  • make this data available on the internet to other, extended-family members (viral marketing!)
  • give the customer two copies of everything on DVD: one copy for home, another copy for the safe-deposit box.

People could:

  • drop off these memories at a local store (kinko's model)
  • ship these memories to a processing center ("warranty" service model)
  • have a processing-van (think huge logo, local branding) come to their home to do the work on-site (van model)

The approach largely depends on how much "distance" from "memories" becomes uncomfortable (for the average customer). Fears could be augmented with strict handling procedures, strict storage procedures, and nationwide branding. Lot's of possibilities here.

Model:

  • Charge customers variable cost (in cents) per item. Outdated video formats would command higher prices. (All costs relative to costs of processing equipment and labor).
  • It will be impossible for customers to sort their items: they cannot possibly filter their memories by level of importance! They will have to give us everything...
  • Customers can come-in with boxes and boxes of items, unsorted, unorganized. We do the work, give them 2 copies of DVD's, other deliverables.
  • How much could the one-time infrastructure costs (per processing center) possibly be? $250k? How many families in suburban communities (like Agoura, Westlake, Thousand Oaks, which has over 15k married dual-parent families with kids). Recurring revenue might be a serious problem. Do digital/analog photo development and other ancillary services? Ahh! Instantly add these newly developed photos to the "archive"...

Like online tutoring services, we would be pleading to emotions: is it really worth risking the loss of your memories? Don't you want to guarantee these possessions for your children, and future generations? ....your parents? ...your children? Threat of fires, floods, tornado's, earthquakes, hurricanes, theft, loss?

What feeling is more innate than the desire to preserve thoughts, memories, emotions? Everybody feels that their life is important. Everybody wants to leave their mark on the world. It is a human motivation as innate as survival and procreation, and ties hand-in-hand with both.

So I got busy (co-founded Academy123) and "parked" the idea. A few years later, I started thinking about it again: putting together an overview, thinking about the workflow, and mocking-up would-be marketing materials.

I proceeded with the assumption that people wouldn't want to part with their photos (I needed to create as little distance as possible between a person and their photos). So, the solution would be to go to their home, with a van, just like a carpet cleaning service.

I ran into a technical challenge: how do I quickly scan all those photos? Varying shapes and sizes. Some in albums. Some stacked in shoeboxes. Basically, I had to assume a huge mess.

My first breakthrough was to think of a "good enough" solution: why not just take a photo of a photo? I could automate the process of capturing the image, cropping it, de-skewing it, etc. In a Ford Econoline van (or similar):

  • Setup a conveyor belt.
  • Have an operator just slap down photos on the conveyor belt, not paying much attention to placement. Can even place entire album pages.
  • At the end of the conveyor belt is a "black box" where the photo enters. Facing down from the top of the black box is a high-resolution camera.  Basically, something in the spirit of this Bookscanner developed by Xerox PARC.
  • The camera just snaps a photo (of the photo!) at regular intervals
  • Later, software (like imagemagick) cleans up the photos: crop, de-skew, separate multiple photos that appear on one album page, etc.
This, I thought, would enable me to scan thousands of photos rather quickly, with a quality level that would be acceptable to most folks.

Later, I realized that companies like Fujitsu make some really high-speed and high-quality scanners. I actually got a hold of a Fujitsu unit, ran some initial tests (scanned a bunch of photos, cleaned them up in an automated way, etc.). This unit has a large auto-feeder (photos) and also a flatbed (album pages?). The scanner is incredible. It basically scans photos just as quickly as a high-speed printer can print color pages. It can do about 60 photos per minute. And, you could just stack the photos any-which-way.

So, I ran with the concept and started fleshing it out a bit:
Firestorm Background

Firestorm - Workflow

Here, I mocked-up a flyer that would describe the service to a consumer. I toyed with the idea of snail-mailing it out to 1,000 households to see the response andgaugedemand (these days, web developers do the same thing by designing a landing page and buying some Google Adwords).
Firestorm - Draft Flyer

I continue to be very much in love with the idea in that I think the need still exists. I think affluent families would pay a one-time service fee of $500 (or similar) to preserve their old photos forever. I think the business could scale just like any typical "franchise" type business (not very interesting when you're accustomed to working on web companies that can scale worldwide with a tiny staff).

Clearly, photos have moved to digital. Successful photo-sharing sites have proven that photos are extremely viral in nature, and I think that these old photos are not an exception. As evidenced by many of the same photo sharing sites, there are a number of ways to monetize the photos long-term: subscription service for storing and backing up, ordering prints, ordering archive dvd's, printed photo products, etc.

I think there might be a short term (10 years?) window of opportunity to "grab" the worlds old photos and bring the old, printed, analog photos to digital. The company that can do this can monetize those photos (and the online viewers of those photos) for many years to come.

Ultimately, I abandoned the idea because it's just not scalable (in the typical sense as described by web developers).

I know that there are a handful of companies out there doing something similar. Digital Pickle has been around for a while, but you have to mail your photos to them (non-starter for mass-market, in my opinion). Kodak had a short-lived ScanVan initiative where they brought photo-scanning trucks to local shopping malls, and I'm not sure what happened there.