Moving from speech recognition to voice recognition

"Hello, computer."

Conversational Design has (almost) arrived. Voice commands as input device are everywhere. People use it to control their smartphone. Amazon Echo is a success. Speaking offers a speed-to-task completion that beats typing.

And right now it is easily undone by a toddler.

That’s my two-year-old daughter, Avery. She is a typical, boisterous and full-of-energy toddler. When I attempt to use voice commands – to draft a text message, search for a Doc McStuffins YouTube video through Apple TV, whatever – she invariably talks over me. Loudly. And then whatever device I’m using fails miserably.

Speech recognition software has gotten much better recently and so we’re starting to realize the benefits of speaking to our devices. We can all speak faster than we can type. This creates efficiency in how we long it takes us to complete a task. But for this to work, it requires a quiet space.

I’ve been thinking a lot lately about how it will look (er, I mean sound) when we are all talking to our devices. It will get quite loud. The workplace will need to change. We will need to have more privacy to work - to speak with and to our computers. This may, or may not be practical (workspace being a pricey expense). The open space floor plans of most offices will be rejiggered out of necessity to accommodate conversational interfaces.

Out of the office - commuting, traveling, at home, wherever - I see two continuing problems with interacting with our devices. The first is societal. We frown on people talking in public, whether it be on the train, in an elevator, or waiting in line. We consider it rude when the individual vocally intrudes on the group.

Working on this post while riding commuting to work. Yes, I get to ride a ferry every day.

The second continuing problem is technology. Our devices need to identify not just the words being spoken, but whom is speaking them. Until they can lock in on the user's voice to the exclusion of all other voices, the interface will continue to fail us. It needs to understand that I'm issuing commands and not be distracted by Avery practicing the Happy Birthday song.

That should be the goal. Getting devices to go from speech recognition to actual voice recognition. When that happens, maybe we'll have something.

On the Essential Sameness that Is Competition

Peter Thiel, Courtesy Getty Images

Just listened to a talk given by Peter Thiel on the “How to Start a Startup” podcast. You can find it on iTunes here. Definitely take the time to listen if you’re interested in starting a company.

Thiel believes that the success of a company is determined by a simple formula:

“For a business to be valuable, two things must be true: you create X dollars of value for the world. You capture Y percent of X.”

X (the size of the market) doesn’t have to necessarily be huge if you capture a large enough percentage of it. He compares the airline industry to online search. Domestic air travel accounts for $195B in revenue annually. Google brings in $50B annually. The search market isn’t much bigger than Google’s share. So which company would you want to be: Google or American Airlines?

 He argues that the goal of every founder/CEO is to establish a monopoly. Do something different from everyone else, go after a niche market and own it. Then expand to related areas, increasing your dominance. Or to put it another way, Amazon started out as an online bookstore. This lead him to another great line: 

“All happy companies are different because they are doing something very unique; all unhappy companies are alike because they fail to escape the essential sameness that is competition.”

He’s repurposing a quote from Leo Tolstoy’s Anna Karenina:

 “All happy families are alike; each unhappy family is unhappy in its own way.” 

I enjoyed Thiel’s talk. He made a lot of great points and expanded his counterintuitive argument that you should not hope for competition but avoid it. Go check it out.

Bad Ad Experiences (part II)

In a recent post I wrote about how poorly designed ads negatively impact the experience users have when visiting a site (and why brands seem to ignore this). Sometimes though, paid advertising can actually undercut a point you’re trying to make.

Slate recently published this post about how Donald Trump’s personal attack on Ted Cruz’s wife signaled a new low for the candidate and the presidential campaign. Its point is pretty clear: attacking Cruz’s wife was purely misogynistic and so far beyond the pale of responsible debate (Slate has written several posts like this about Trump, such as this one that states his only consistent position is misogyny).

And then you get to the bottom of the post.

In the same post is a “promoted story” from LifeDaily entitled “20 Things You Should Know About Donald Trump’s Wife.” Kind of undercuts that point Michelle Goldberg and Slate are trying to make. 

The Trump’s wife “story” is obviously clickbait, so I’m not including a link to it. But Slate is taking the money from Outbrain (the company responsible for serving up the promoted stories). It’s a bit jarring for users to read a post about how wrong it is to gossip and attack a candidate’s wife, and then see an ad on the same site that is doing the same thing to the offending candidate’s wife.

Take care what ads you serve up content sites.

Bad (Ad) Experiences

I’ve been using the Yahoo Weather! app for a while now. It leverages flickr to serve up beautiful geo-tagged photographs that match the location and (roughly) the current weather. It’s beautiful and evocative.

Beautiful, right? And then you scroll down and see this ad:

Not so beautiful. I have always wondered why so many companies rarely consider the ad that’s placed in an app. They do realize that the user experiences the ad along with the rest of the page and doesn’t separate the two? Or to state this more clearly:

Users experience the ads you place on your site along with the rest of your site. They don’t separate the experience of one from the other.

This is intuitive, or at least it should be. Bad ad experiences contribute negatively to how users feel about the sites serving them up. It would seem that it is in the site’s best interest to improve the ads and thus improve the overall site experience.

Or (try) to take a look at The Atlantic.

Basically you can see the site navigation (this is actually an article detail page) and that’s about it. Ads everywhere else.  The Atlantic is a highly regarded publication with great writing and high editorial standards. Someone spent a lot of money designing a website to match the style and caliber of the print magazine. And then they mucked it all up with ads that have little connection to the magazine’s brand.

I’m not against advertising or the need to place ads on sites and apps that offer free content. But it seems strange that a company doesn’t care how off-style an ad might be placed within their otherwise impeccably designed site. It’s jarring.

What I’m asking for is that companies avoid the garish. And there’s certainly a way to do it. Take the biggest provider of advertising on their site: Google. Ads are everywhere. But the ads don’t intrude and look out of place.

Ads are necessary and important (they help keep the lights on, and content free). But this shouldn’t be at the expense of the user’s enjoyment of a site.

Please: consider the design and style of the ads you place on your site.

Conversational Design: An App(s) Review

I recently came across a couple of new apps that utilize conversational design. Each focused on solving a different need. So I decided to experiment with them and share what I found. The apps are:

• QUARTZ -  a news app:   iOS  |  (Android is being developed)
• HOUND - a digital assistant:   iOS  |  Android

QUARTZ

Quartz mobile is the new app from the same-named news site. The interface replicates a chat window. When the user opens the app, headlines appear as new messages and after each one, the user can opt to learn more or see what’s next (it also often employs emoji in the “learn more” choice).

The nod to conversational design is obvious, and also fake. Currently, the user cannot ask Quartz for information on a particular topic. You can’t ask Quartz, “Tell me what’s breaking news in Boston?” for example. Quartz is effectively a slow scrolling newsfeed, without opportunity to either direct or personalize it. This is release 1, so it’s possible that functionality is coming (it becomes immediately more interesting if it does). It’s possible that it is learning more about the user based on the headlines they show interest in, but it is not immediately evident.

You can watch a TechCrunch review of Quartz here. 

 

HOUND

Hound is meant to be a Siri competitor. It’s onboarding is effective, asking the user to say specific questions, each one demonstrating a different capability of the app. I suspect that at the same time it is calibrating the app to the user’s speaking pattern. I’m not sure of this, but if it is a brilliant approach to teaching the user how to use the app while also making app setup simple.

If you want to learn about the current weather – either in your current location or somewhere else entirely, this app works smoothly. Contact integration is impressive. Saying a contact’s name brings that information up immediately. Saying “Text [contact name]” also works as expected. It does not however support email integration.

Hound also integrates iTunes so requesting a song usually results in success (if you don’t own it, you’ll get only the free sample through the service). It also has Uber integration.

I found voice recognition worked well, better than Siri. Where I struggled was not knowing all of the possible commands for Hound. When it doesn’t recognize a specific trigger in your question, the result is a google search for the exact phrase you spoke. While this can certainly be frustrating, it’s worth noting that in all instances it searched for the exact phrase or sentence I uttered. In other words, it’s speech recognition worked well. I’d like to see additional functionality added to the app, such as shopping. Hound will likely continue to grow in features as they improve the app.

Read a TechCrunch review of Hound here.

To start (a company)

I want to start my own company. I want to develop a product based on my own BIG IDEA, form a company around it, hire people to my cause. I want build something. 

I think about this often. There are several ideas for products, all somewhat related to a general theme. Not necessarily a specific BIG IDEA, but something or several things.

I don’t have any cofounders to help me. I certainly don’t have any tech founder friends to lean on and write code. My skill is lacking here. I recognize that as a definite obstacle. I do have a mental list of people I would like to ask.

Dilbert, by Scott Adams. Thanks to @lipmanb for tweeting it.

I really do think often about starting my own company. I don’t think this makes me different from a lot of people. Rather, I suspect a LOT of people out there daydream similarly to me. Beyond the obstacles I tongue-in-cheek outlined above, I consider the real risks of taking such a chance:

 

CHILDREN

I have two daughters – Avery who’s 2 and Lucy who’s just 6 months. Coming home to them at night is amazing, and exhausting. They command my full attention before and after work, and on the weekends. How could I devote the often stated 100% focus to my potential startup while maintaining my commitment to my family? I actually took heart in a recent post by Rob Gonzalez, CEO and cofounder of Salsify, where he talks about the balance between founding a startup and personal life. Interestingly he talks about the advantage of parents in this environment – they have focus because they need to be. This theme was similarly espoused by Paul Graham of Y Combinator in his talk "The Counterintuitive Parts of Startups, and How to Have Ideas." Listen to the Podcast. But let’s be honest: I’ve never started a startup and everyone always talks about how much time must be devoted to it. Am I just cherry-picking the the advice I want to hear?

A GOOD JOB

I have a good job and get paid well for it. There is some reward in the work I do. Kids are expensive. How can I responsibly risk my income on such an unknowable endeavor? Losing my income for however long would put a huge burden on our family. And who knows when I would start to take in any pay, let alone match my current salary. If this is an ego – or selfish – leap to indulge personal wishes I will have failed in my primary responsibility as husband and father.

 

NOT IN MY twenties

I’ve been out of my 20s for a while. That’s when you’re supposed to try your hand at starting a company, right? Sure I have more experience now than when I was younger, but I have more responsibilities (see reasons above). What would I be giving up to make this happen? If it doesn’t work out (and the statistics say a startup is likely to fail), what damage might I be doing to my professional career? Am I setting myself back a few rungs? I don’t have an answer to this.

---

There are many reasons why starting my own company is a bad idea. Again, this realization makes me like many, many other people. We all want to be the boss. We all want to do great things. We see only success, and not failure. With all of that, I still hope to start a company one day. I don’t know when that will be. And I don’t know what it will be. Obviously I don’t know if it will be, but I’m hopeful.

Maybe as a dream, it can never fail.

Maybe.

We're Going to Be Alright

Today is Super Tuesday. Election day! This morning I dropped my daughters off at daycare and then went to vote. I took a later boat (I commute to work on a ferry – it’s awesome). While waiting to leave the dock, a lady sat down next to me: “Oh look, you’ve already voted! I need to do that.” She had seen my “I Voted Today!” sticker on my jacket.

We talked about whether there was a wait at the polls (there wasn’t) and what time they closed (we thought it was 7pm; turns out it’s 8pm). She was excited about doing her civic duty, and said she would vote later that night.

Once in the city I headed to Starbucks, as I do every morning. Stepping up to the barista, a big smile came over her face: “You just reminded me!” she spread her arms in a happy, excited manner, “I almost forgot; need to go vote today.”

She then took my order – a grande iced coffee and a blueberry muffin. She comped me the muffin (thank you!).

Watching the news we tend to get the impression that all things are dire and serious foul things are afoot. The other side hates us and will lead us to ruin. Conservatives and Liberals should just form their own countries away from each other. It’s pretty dire.

It struck me today how happy people were to be voting. No one asked “Whom are you voting for?” No one tried to influence anyone else. I felt a genuine sense of community among everyone I interacted with today. This is not the feeling one gets following election coverage.

I have nothing profound to say about all of this. I just wanted to note that I felt kinship with my fellow Americans today and that is a good thing. Despite all the doom and gloom that always leads, I think the country is filled with good, well-meaning people. We’re going to be alright.

If you haven’t yet, go vote. 

Coin and 10x Better

Last weekend I was out to dinner with my wife. An older couple at the table next to us was brought their bill and when the waitress returned, the husband pulled out his wallet.

From his wallet he pulled out two credit cards – one a Coin card and the other a regular credit card. The gentleman explained that he would like her to try the Coin card first, but that he also anticipated problems.

“The Coin card’s great, because it allows me to store all of my credit cards on it," he said, "Problem is it doesn’t work all the time. That’s why I also carry this card. If the Coin card doesn’t work, just use this.”

This was the first time I had seen someone attempt to use Coin at a restaurant. I find it odd that Coin hasn’t become bigger than it is and at the same time, the conversation I overheard pretty much explains why it hasn’t gotten bigger.

It’s an object lesson in the rule of 10x better products. Coin sounds like a great idea, but users haven’t made the switch in compelling numbers (note that the gentleman had to explain what it was to the waitress who handles credit cards all day at work). Coin started shipping a year ago, in early 2015 (there was a Beta available before that). I suspect there are several problems:

 

INFORMATION

Users are unsure of trusting this new device with all of their credit card information. Perhaps in users’ minds it represents a single point of failure for their financial security (if someone steals their Coin card, then they have ALL of their credit cards). It’s on Coin to do everything they can to inform customers and convince them that it’s safe, and easy to use. Here is Coin’s FAQ. Judge for yourself if this allays users’ fears or not.

 

REALLY BETTER?

I carry two credit cards (one of which I never use) and an ATM card in my wallet. I would like to lighten my wallet, sure. But are the two extra cards in my wallet such an inconvenience that I need a new solution? It’s perhaps not 10x better.

 

RELIABILITY

When Coin launched, a friend at work ordered one. He was quickly frustrated with how often it failed to work as designed. He quickly gave it up after about a week. Apparently this is a problem that persists a year after launch. This must have a huge impact on user conversion. 

It’s a reminder about how hard it is to launch a successful product. A good idea isn’t enough. You need to be solving a real problem and be a far better solution. You have to educate potential users on why your solution is so good (don’t take it as an article of fact that they’ll understand intuitively). And you better execute flawlessly. 

The waitress returned a few minutes later and explained the Coin card wouldn’t process; she ran the backup card as a result.

On Conversational Design

One thing that has me really excited about new products is the growing trend of Conversational Design.

I recently wrote about the breaking down of siloed experiences created by apps so that contextually related tasks can be combined into a single experience (e.g. ordering an Uber from Facebook Messenger). This is partly a result of the user interface transforming into a dialog between user and application. Text being the great equalizer, if a conversation can happen, all things are understood and possible.

This trend will likely be hastened by the advent of two other technologies that are quickly gaining traction: Artificial Intelligence and Virtual Reality.

AI as it relates to conversational design is obvious. AI involves a self-aware machine that can learn from interaction. Conversational Design doesn’t require AI – when I “chat” with Slackbot to complete my user profile, it’s not learning; rather the interface is cleverly constructed to appear to be a conversation. No doubt though, as AI is improved and integrated in everyday products, the net effect will be an increasing use of conversational design in all things.

Virtual Reality, I believe, will also help spur on conversational design. Traditional user interfaces will seem to be clunky virtual reality. They would also threaten to break the illusion VR is creating (I’m in this amazing, majestic landscape; let me pull out a keyboard to navigate elsewhere). Dialog seems to be a more fitting UI in VR. Hence, conversational design.

And this brings me back to the app experience today and why it needs to be replaced. I suspect that a user will not want to have conversations with numerous different applications, each with different responses and cues to natural language conversation. Users don’t want to “train” 20 different bots to understand them. They want a single voice. They want a personal concierge. 

Apple has Siri and Amazon recently launched Alexa. Each will get better with iteration (I mean, Siri can’t get worse, can it?). But I think there’s a risk in these platforms being built by existing businesses. What if I don’t want to own an Apple product? What if I want to shop somewhere other than Amazon? I believe there’s an opportunity for a new company to build a conversational product that is agnostic to brand and business. Something that handles one's Twitter, Facebook, and Instagram feeds, while allowing users to request services, such as an Urber or food ordering, and allows shopping across all online retail. This digital concierge will work everywhere.

That seems like something people would be interested in.

A Blog About Product

In 2013 I was working with my team building a digital product. This was a little unusual because I didn't work at a product company.

A little background: I've spent the past decade working in UX, specifically for digital agencies helping big-brand companies design marketing and ecommerce websites. Agencies are really good at building something to specifications provided by a client. But we'll never be mistaken for a product company (even though more and more agencies are trying to pull this off).

The product we were trying to build was a geo-location based social media app whereby users were able to "place" digital content in the real world to be discovered by other users walking by. We failed miserably. But this post isn't about that experience.

I bring it up only because it was at this time - while I was trying to figure out what we were doing wrong, or trying to figure out how to do it right - that I came across this Medium post: The Secret to Designing the Right Product by Matt Schlicht.

This post was eye-opening to me, introducing me to a new way of thinking about design and web development. It was a very non-agency approach. From this post I began reading everything I could find on the subject of product design and lean methodologies. The subject invigorated me and thankfully there's a lot of content out there on it.

This is why I'm starting this blog. I'm interested in building products (and companies) - how to do it well, and new ways of working. I intend to use this blog to talk about what's happening in the space, as well as comment on what I find interesting about those developments.

If you're interested in this theme as well, check back. I haven't settled on a cadence, but I'm hoping to post at least a couple times a week. And reach out to me on Twitter and LinkedIn.

More to come.