why shipping frequently is important?

Yes, it is important to ship often, get feedback, and make the software better. Especially if you’re developing a service shipping to customers, it is a great way to get the code executed. Shipping improves quality.

It’s given that first couple times, there will be failures, more bugs, fire-drills that would make you look bad in front of your management and customers. Especially if you’re transitioning from longer release cycles. But it will be a short period of time, and benefits will show themselves pretty quickly.

All good bugs are found when the service goes live and running in front of customers with their data. Those bugs cannot be found in the lab, while testing. A test cannot match to a customer’s mind or behavior, and the test infrastructure cannot match to the combined force and configuration of phones, tablets and computers out there!

How do you know you’re ready to ship?

Most often, you won’t. You’ll feel ready, with an expectation of a learning opportunity coming up. You can increase the confidence level, by having regular automation tests, component level tests, scenario level tests (for defined scenarios). And these will cover a lot of paths in the code, but not all unfortunately, and enough if you had just these.

The code can tell you more with its data. Data that comes out from your service, telling you how much the failure rate is per scenario, and where the failure is, and provide enough information in numbers so that you’ll be able to fix it in a quick short amount of time. Logs may not be be as affective as is development stage. Real usage would be high, and there will be a good amount of logs collected due to the usage. Tracing where the failure is and debugging live would be a big time spent for too small.

So if you have a habit of shipping often, preferably in small chunks and gather data about how things are going from the code that goes out in front of the users, the quality of software increases. Because the increment goes out is tested by real users, running with real data and if a bug comes up, it will be fixed rather quickly because the change went in the other day is probably small.

IMG_5035

online holiday sales by numbers

IBM did another benchmark this holiday season. Quite interesting for a data guy… They tracked online sales for black friday, cyber monday and Christmas Day for the US, and Boxing Day for the UK.

Online sales are growing, compared to the last year, which sounds good, and seems like economy is getting better. Compared to 2012, black friday and cyber monday sales grew around ~20%, and Christmas Day sales grew 16.5%. Boxing Day sales did grow 40%. Mobile devices are becoming more integral part of our lives as they had a growing pie in online sales.

We increasingly use mobile devices, to search and read about the merchandise we are interested in. Overall 30% to 40% of online sales traffic originated from mobile devices.  Smartphones were used twice as much as tablets.

mobile_traffic

But when it comes to turning traffic into sales, tablets drove the purchasing, and for each sale happened on smartphones, there were two sales happened on tablets.

mobile_sales

This could be due to the push promotions happened throughout this period. Per the report, retailers sent 37% more push notifications to their users throughout this period.

One explanation of this could be that people receive these notifications on their smartphones, read and learn on their phone as much as possible, and buy later from their tablets. There is also a convenience factor of tablets as rendering and execution of web pages are much better on tablets, and putting in credit card number via a tablet is much easier. Often the layout of websites are not optimized for viewing on a phone and that makes it harder to use. This also could be a perception issue as well, because tablets feel more like computing devices.

The reports have also a section for social influence. Overall, Facebook referrals converted to sales more than Pinterest referrals, but average order value from Facebook was lower than Pinterest. For example on Black Friday, for each Pinterest referred sale, there were four Facebook referred sale, but Pinterest shoppers spent 77% more per order.

All nice numbers, and sales are growing…

So if I were to operate a web retail shop, I’d optimize smartphone interface in such a way that one click sales would work, without much effort from the user. Give as much info as possible, with minimum overload (it’s a small screen, do not get the user exhausted). Of course mobile app makes a lot of sense, but on the way there, there is a browser on each device now, making use of it also makes a lot of sense.

I’d use Pinterest in such a way that you get your items pinned, it’s a window shopping experience, people like what they see, they buy. But recommendation is still big, and will be big for a long time. Get the recommendations, comments, feedback published on Facebook. Get people talk about your stuff on Facebook, so that their friends buy through Facebook. Given that Facebook is becoming the place where the people of purchase power hangout (moms and dads), they would be curious about the details and read more…

And I’d also invest in measuring the numbers, because I’d like to know what people are doing on my website. I’d like to know about them, and learn from them.

WP_20131228_002

day twelve: mostly doing web and photo editing

Christmas week. With our family friends, we went to a close by hotel just on the other side of Snoqualmie Pass. I was expecting some snow, but there were none at the hotel. With a small vacation, and work being slow due to the Holidays, I did spend some time with MacBook.

I’m mostly on the web, using either Safari or Chrome. If I’m not on either of these I’m using Lightroom. I took quite a bit of photos over the last couple days, and Lightroom has been fun doing light touch ups. I like learning new skills, and Lightroom is a real and serious uncharted territory, and I have a lot to learn and practice. Some recent photos are here.

This morning I found “Activity Monitor” of Mac. It shows which apps I’m using and their energy impact on the laptop. A screenshot from this morning.

apps and power consumption

Energy impact is a relative measure based on many parameters including CPU usage, network traffic and disk activity. Average energy impact is an average over 8 hours.

Obviously, Lightroom requires relatively more energy compared to Safari or SkyDrive or any other app. I guess Safari rendering takes some power. Also need to say that, even though I play with Lightroom, and it’s up quite often, Safari is up and running almost all the time (ditched Chrome lately). Apparently OS X Mavericks  puts apps to sleep via App Nap to conserve energy. Safari seems to be doing the same thing to web pages in tabs.

Feels like the OS does everything in its power to conserve energy. Maybe a little bit excessive and aggressively.  So far did not feel any sluggishness of apps waking up from their nap. Skype, for example, comes up to the foreground very quickly.

screenshot

day four: Office

After the accident yesterday, I re-setup the Mac. Then went to Office 365 website and downloaded Office 2011 for Mac. I needed to work with Excel to create some charts. Installation was seamless, and quick, and was greeted with a very well known UI: Excel.

So I can have a convenient, and known experience in this unknown territory. I felt good.

I also started playing with spaces. I used to have 4 virtual desktops back in college while running Linux. My window manager was Blackbox for a long time, and used to use workspaces very effectively. At that time, I was very much doing a lot of things at the same time. Nowadays, I try to do one thing at a time and not looking at other things, as much as possible. However Spaces could be helpful in organizing the desktop, and reduce clutter.

Between installs, setup and meetings, I did enjoy sometime working with Lightroom.

  

build. learn. iterate.