The One Thing

I was reading “The One Thing” by Gary Keller and Jay Papasan lately. The book is an Amazon best seller, WSJ best seller, and has over 2000 reviews with an average rating of 4.7 out of 5 on Amazon. It is a great book. Some of key learning points that I wan to pencil down here:

  • Multi-tasking is a lie – I see at home and at work, many people including myself do multi-tasking – seemingly very busy and feel good about it, probably because we think we get things done at the same time – so I am “productive”. That is NOT right way to be productive. One Thing at a time. No multi-tasking.
  • Always ask the focusing question: What is the one thing I can do such that by doing it everything else will be easier or unnecessary? This shall be applied to many aspects of one’s life:
  • Time blocking – use your calendar to block the time – 4 hours a day suggested by the authors – to focus on the one thing. Everything else will be a distraction. Even if other things are getting into your way, you have to protect your time block. In case you have to switch context, re-block your calendar. Four steps of time blocking: 1) time block your time off; 2) time block your ONE thing; 3) time block your planning time; 4) Protect your time block.
  • Be a maker (do or create) in the morning, manager (oversee or direct) in the afternoon, as morning is probably the best time to focus.
  • Four thieves of productivity: inability to say “No”, fear of chaos, poor health habit, environment that does not support your goals. The first two thieves are common to me. My takeaway from this is that if the thing does not connect to my ONE thing, just reject it, deflect it, or let it be.

Overall this is a good read. The authors even built a website that offers videos and tools to help the readers put the ideas in action.

Google My Business and Messaging

Here is my experience (so far) trying Google My Business.

I can easily add my small gig (a consulting service) to the listing. Also, GMB provides a simple website builder for my business. This was instantly published as I can see it from Google Maps search on my phone:

Notice the “Message” button on the bottom…

So yes, my business is listed. If someone hits the “Message” button, a screen will be launched on the phone to allow her/him to send messages to me – I actually will receive the message from within my Google My Business mobile app, where I can reply to those questions.

Turns out this “Message” feature for a business is provided by Google Play Service – so anyone who does a Google search or Google Maps local search will be able to send messages to a listed business from Google search or Google Maps, without requiring a messaging app (Google calls this App Preview Messaging).

I think it should have been designed in a better way, where that someone can actually use her favorite messaging app to communicate with a business via SMS/MMS/RCS, AND has the option to disclose or hide her identity.

Two advantages of this approach: 1) a user can manage such conversations in a single place – the messaging app; 2) business will have a better chance to engage with customers in a persistent conversation, instead of the transient window which will only show up when you hit the “message” button of that business again.

Disadvantages are also obvious: business needs a real phone number to communicate with a user. And, the conversation will be between two phones then. Compared to current design, which is basically app-to-app messaging through Google so Google can offer message analytics (“insight”) for business.

That is my first 2 hours with GMB.

Small Biz Online Presence and Opportunities

A typical customer journey with a small biz can be as simple as the following:

Discover –> Visit –> Browse and Ask –> Order/Purchase –> Customer Service 


For a small biz, now in order to be discovered by prospective customers, there are a few channels you want to consider (in addition to online advertising):

  1. Google My Business listing: this is what Google Maps search or local search come in to find you. You want to make suer you list your biz there and manage it by using the Google My Business app. This seems to be also used by Google Assistant. GMB also provide some API that allows programmable access. Moreover, there is an interesting “Messaging” button for your GMB that will allow one to send you a message to your GMB app.
  2. Yelp listing: you need to have a page on Yelp. Amazon uses this listing, along with Yext. API is available.
  3. Yext listing: a “knowledge network” which basically helps biz manage your online presence. Amazon Alexa uses Yext for voice local search. You can also use Yext service to manage your biz on multiple channels.
  4. Bing listing: this is used by Microsoft Bing and Cortana.
  5. Apple Business Connect: Siri uses this listing.

Of course, search engine optimization is also a key when it comes to be discovered through user’s web search.


Apart from these general listings, small biz may have its own website or a Facebook Page. Most eCommerce platforms already provide services for website design and hosting (with shopping cart support), and Facebook Page creation and management, as well as other social networking eCommerce channels such as Shopify over Pinterest.

What happens when a visitor is on your site and has some questions? Fill out a form? Well there is a better way to generate leads from these questions: use a 24×7 chatbot that can not only answer questions of your biz, but also can schedule meetings/calls between the customer and your sales team. Same for your FB page visitor. So there is an opportunity here.

In summary, for a small biz, with limited budget for online marketing, there is a need for a low-cost, easy-to-use, one-stop service that allows them to manage listings across channels, capture visitor/customer questions (in text or scripted voice call) and convert to sales lead, and integrate this data into CRM.

Specifically, for voice based conversational eCommerce, there are 3 levels:

  • level 1: IVR – interactive voice response. “Press 1 for product listing, Press 2 to talk to an agent”. Very basic form of voice service.
  • level 2: Simple question/answer – this is the voice version of a simple text bot. It is programmed to provide basic information of the business, with some natural language understanding and natural language generation.
  • level 3: full-duplex voice assistant – this is close to a human assistant, instead of being able to easily recognized as generated voice.

Status quo of the industry is that everyone can do level 1 IVR; and more brands are exploring level 2 (as mostly manifested on Alexa and Google Assistant). Level 3 is still at early stage although Google and Microsoft demonstrated some recent developments.

Mobile Advertising Round-Up

Some people say it is a big challenge to map web user to a mobile user (and her context). On the Web, it is already common practice to figure out a user’s behaviors and do target advertising. But how this is mapped to mobile?

Cookie Syncing solves the problem of mapping users across different sites, by Ad exchange and DMPs.

Rule based and Machine Learning based – Rule based programming does not solve large scale and high dimensional problems. ML/DL can do.

User data is not being sold from to facebook. Instead, Costco provides an ads slot to Facebook, and Facebook decides relevant ads to be shown here. To do so, Facebook needs to know the user’s facebook ID.

Take a facebook chatbot for example. The chatbot (of a brand) knows the facebook user’s profile, and can thus provide some target information in the conversation with the user.





Mobile AI – The Big Picture

There has been a lot of media coverage around “Mobile AI”, from various parties in the value chain. By definition, mobile AI encompasses cloud based processing plus on-device computation. Depending on the use case, the model of Cloud v.s. On-Device can vary. Here device can be a phone, a smart home hub, a TV, a module in a self-driving car, or anything we call Edge for that matter.

What’s happening now?

  1. Chip makers – let’s make some AI chips to take on specific computations on a device. Example: Apple Neural Engine as part of A11 bionic chip, Huawei has an NPU in its Kirin 970. Or, the SoC itself can already support some AI computing tasks, for example Qualcomm Snapdragon NPE. 
  2. OS platform provider – our next version of the OS will have neural network support from ground up. Not only our apps and services are doing some AI things, we are also offering tools and SDK for third party to leverage the newly added ML/DL framework to achieve your own goals. Example: Android TensorFlow Lite; Apple Core ML;
  3. Device OEMs: they have unique opportunity to leverage AI chip powered computing capabilities, plus data only OEMs can have, to deliver AI to users as well as 3rd party developers. Example: Huawei has HiAI open to public.

What are the top challenges in this space? Or put it in another way, what are the opportunities?

What value does mobile AI bring to user, or broadly to the ecosystem? We see AI for making the camera smarter, saving battery life, system-level scheduling for performance improvement, application specific AI use cases – Good examples are Google Photos allowing you to search by text for an object, and Google Keyboard GBoard leveraging some sort of transfer learning to continuously optimizing the prediction model in the cloud without sacrificing user’s privacy (user’s input).  Anything else?

There are at least two domains in this spaces: one is to enhance the applications, services, and the system – I call this domain “AI-enhanced apps/services; the other to is create new opportunities, cross device, between services, with some sort of “link” between them such that a new value chain is established. I call this domain “AI-created new business”.  More to come in this space…