Why Retail is Failing

Retailers keep closing shops. This has been the trend of 2017. Often it is being blamed on Amazon and called the Amazon effect, but that’s really an incorrect attribution of the cause. There is a very different and very fundamental reason for this breakdown in retail. That issue is that most retailers have not understood how to make money in a digital economy.

Retailers have always been masters of making money on physical products. That is what they do and they often do it well. Those principles which drive them are rooted in what Adam Smith first wrote hundreds of years ago about how trade and economies work. And those principles have worked quite well for a long time. However, as digital products grow, we are seeing that many companies are struggle to make money in a digital economy.

Digital products are very different from physical ones. Yes, they are often very similar but the barriers to entry and means of production change the game quite a bit. A digital product has very different marginal costs. Physical products were all about getting the means of production down lower than the competition. The per unit cost was very critical, but with digital products, the per unit production costs is almost nothing every time. When a physical book is created, in order for an author to share that book with customers, a new book is often needed, which means it must be produced and that cost is then passed on to the customer along with a profit margin. Now an author can create an ebook and once made, the per unit is nothing. That book can be shared with 1 million customers or just 1,000 customers, the cost per book never changes. This new shift in cost per unit is one of the things that is impacting retailer’s abilities to compete, they haven’t grasp how to deal with this change.

Add to that, the change in social behavior. Up until about 10 years ago, the rite of passage for many teens America was to get a driver’s license. It was one of the most important events in a teens life. But now teens are waiting until they are young adults in their twenties to get a license. Why? There are three main reasons we drive in America and in many modern countries:
1: To get to work.
2: To get products we want to consumer.
3: To meet people we want to communicate with.

Many teens don’t work so they don’t need a car. Add to it, the means of consumption were now available online. When I was growing up you could get a pizza delivered at home. Now, you can get pretty much anything delivered. And last, most teens prefer to communicate online, not in person. This kind of shift is important because if people no longer want to go to a physical location to socialize and consumer, then you don’t need a physical store. Many retailers are trying to fight this trend instead of embracing it.

If retailers want to still be around, they need to focus more on these kinds of trends instead of fight the old pricing wars they are used to. The digital economy has its own rules, rules that are very different at times from the economic rules we have played by for decades. Understanding and embracing the news rules are the keys to survival.

Analysis of Linkedin in Connection Data

Analysis of Linkedin Connection Data

Recently I downloaded my Linkedin data to see what it might show me. I’m still waiting for all the data but Linkedin gave me some data very quickly. You can do two types of data pulls, a quick one which gives you dates of when you connected to people and a much larger one with more information, which is what I’m waiting for. Most of what they gave me was not that useful for the analytics I wanted to do. But I did get some interest bits out of it. I joined Linkedin on April 14, 2004. Linkedin itself launched in May of 2003.

It was a quiet place back then when I joined and very different from what we have today. Back then you couldn’t post anything, there weren’t groups and you were supposed to only connect with people you know in person. There was a Yahoo group where people talked about Linkedin and I will say that Linkedin was actually easy to talk to back then. An actual person who worked there would frequent the forum and talk to everyone, answering our questions. Of course, now it seems they do what they can not to talk to us. In Linkedin’s own words, “it was slow going in those days with some-days only 20 people signing up.” Although they had some features I liked such as ranking. You could see who was the top ranked person in your metro area or field, things I wish they would bring back. But one thing has never changed with Linkedin, they were then and still are, slow to make changes by listening to customers. Linkedin’s best features were always customers driven and Linkedin’s worst changes seem to be the opposite.

I was never a huge invitation person, I don’t send that many out. Being an early adopter has its benefits. At one point I was the most connected person in the Twin Cities on Linkedin. Then the recruiters discovered Linkedin and that changed really quickly. That’s about the time Linkedin changed to only showing you 500+ if someone had more than 500 connections. Which also had its benefits. People who understood the numbers game of Linkedin, would connect with anyone who had more than 500. That was me, I had a steady flow of connections coming to me. And then with career changes I made, being at the forefront of Big data and Data science, I have had a steady flow of connections coming my way. Much of my growth on here is very organic thanks to variables that were somewhat out of my control. The recruiters came and brought with them lots of new people who wanted to connect with them and then the rise in the fields I work it did the same.


Analyzing the Data

With the first packet of data I was sent, I saw one of the first things I could easily do was analyze my connections. I wanted to see how many connections were added per year and the total growth to see if I could find a pattern. With 10 plus years of data, I had plenty of see any possible patterns.

My first chart shows the growth of new connections added each year. As you can see it was a strong growth period from 2005 to 2008 and then a rapid drop that took years to recover and even then, I never saw the growth I did in 2008. I wondered why that was, what was going on that impacted the connection numbers. Since I always like to see what external factors were going on, I looked at economic data and focused in on unemployment. We saw a massive spike in unemployment from 2008 to 2009 and a slow recovery after that. Could that be the reason for the change in numbers? Possibly. With fewer people working, fewer people may want to connect. Also, the data set I was first given doesn’t have location so I can’t tell how many people are from which country but I’m going to assume most are from the US or an economy closely tied to US economic activity. I make this assumption since my posts tend to get a lot of US based traffic which leads me to assume the bulk of my connections are US based. If left alone without trying to make connections, can Linkedin connections give us an indication of anything? I don’t know, 2016 was a good year for me with over 700 connections, the best year since 2008. This year looks like it is on track to come in close to that number as well. For me, this year has been up and down. Trying a new business can add connections but doesn’t always translate to more money. Whereas 2015 and 2016 were good years for me yet the difference in connections was significant. Although there may be some correlation, I would say the correlation is fairly weak with a number of other variables coming into play that still need to be identified.

You can see in the total growth that things seem to be fairly steady from year to year with 2008 and 2016 being the outliers. I would need more data to get a better understanding of what is going on. I may try and break things down by month to see if there is a seasonal effect on connections. Location would be nice to have as well to see if there is saturation or dispersion of connections over time. When I get the full data set from Linkedin I will see if that data is there to make those kinds of analysis. For now, it is interesting but not at a level we can make any firm conclusion about. There are trends that appear to be there but the cause is still very much a theory.

Let me know what you think of the data and any next steps I should take.