If you are active in merchandising you’ve surely heard already about the Long Tail. Chances are you’ve seen the graph as well…

It’s all about the distribution in popularity of the products in your catalogue. At any moment in time about 20% of your products are highly popular, and about 80% not-so-popular - but, well, they are selling firmly every now and then.

The 20-80 numbers are not set in stone and the debate around them is hot. But they give you an idea that a lot can happen at the extremes, not in the middle: Either very-popular, or rather unpopular and probably niched, selling in small numbers.

Is the Long Tail static?

What bothers me in the graph below is its static appearance. You don’t get a clue of the effect over time. How does the popularity distribution for these products change over time? Furthermore:

  • Should I care?
  • Could one influence this change?
  • Does it always need to look like the graph below?
  • How should a Long Tail ideally look if you are to increase incomes?
  • What are we, ecommerce people, able to do in this not-so-easy landscape?

In this blogpost I will try to give some answers to these questions.

Day 1

Long Tail Apptus algorithm - Day 1-5Let’s go back to the graph, and suppose we assign it to Day 1. (For simplicity let’s consider the popularity distribution constant over one day. Day numbers are also used mainly for exemplification).

So on Day 1, what your online shop sells is:

● Certain trending products (red): these are just a few products with very high popularity. They are your best sellers, your hits. Attention: These hits have a short life in the red, but they are hits (at least) for Day 1.

● Many other products just sell in a stable consistent way, but not spectacularly (yellow). And yes, the Long Tail is to a great extent static.

Now, some products may not sell at all, but they are not shown here -- they fall on the right side of the diagram, due to popularity close to nil.

Day 5, Day 9: What happens to the trendy products?

Let’s now jump in time over few days. Say we look at the same products in the plot for Day 1, now on Day 5, and on Day 9. Short story: The Head partly vanishes, partly dissolving into the Long Tail…

Long Tail Apptus algorithm - Day 5 and 9

As days pass, what you now sell from the old bunch includes:

• Fewer and fewer of the previous trending products are still trending (red)
• Previous trendy products move more and more in the yellow region
• A part of the trendy will drop off completely from the yellow region
• The stable products (yellow) will keep on selling consistently over time – no remarkable change here.

New products will arrive in the trendy region, not marked here on the diagram. (If you’re wondering, they will pop up in the white spaces of the old red region).

Day 14: When the trends of previous weeks are gone

Long Tail Apptus algorithm - Day 14Two weeks on, the graph (excluding new trendy products) looks like this:

• Perhaps only one or two (let’s say) of the products trending two weeks ago

• Some of the old ’hot’ will move to the stable sellers/yellow range and continue to ‘live’ there

• The yellow range stays almost unchanged (just a few newcomers from the red/hot). Its stability should be notable for you: You need to consider whether and how to exploit it!


Can I sell more of the Long Tail?

This is equivalent to lifting the curve, because yes - the abrupt drop away is not good for sales. Now, whatever you do to lift the popularity, it will cost. So of course, you cannot do this for all the products - you need to be selective.

Then you may want to consider mainly the intermediate region, i.e. products with mid-range popularity. How do you do this?

Long Tail Apptus algorithm - lift

There are typically two non-exclusive ways:

1) by lowering the price of these products;

2) by exposing more of these products.

As the spectrum of products in your popularity distribution curve varies continuously, and you normally handle a wide product catalogue, you should not design the campaign manually. After all, humans are particularly good at making biased selections.

Artificial intelligence, as opposed to natural (human) intelligence can aggregate the data in an automatic and systematic way.

There are other important elements to ponder here. You’d probably want to weigh in the price and margin as well, to concentrate on the business objective that matters most for you. For this it is even more important to use AI to cleverly integrate your data in the campaign design.

So you need a system continuously ranking the products and selecting what to expose in campaigns, all while timing it right for each and every particular product you target with the campaign.

What’s in the Long Tail for me, the merchandiser?

What you seek as a merchandiser or e-commerce director is to be able to expose both the right trendy products and the right not-so-trendy-but-stable-in-sales products - but to do so at any moment in time, and in an automatic and intelligent way (avoid human error!).

Because if you miss any of them, you’ll lose sales, right?

The make-up of the popularity curve varies fast over time. It varies from hour to hour, along with the small or large trends, and it moves over years, together with the field dynamics. All the same, some sectors claim that the long tail is dead, that the top 1% accounts for 75% of revenues, or even 99% [1-3].

Since the publishing in 2006 of a best selling book [3] there has been an ongoing debate on the actual effect of the Long Tail phenomenon in eCommerce, with a particular focus on goods that can be downloaded digitally (media), but lately also for other areas - for example fashion [6-8].

On one side there are aspects of the popular Long Tail theory of Chris Anderson: “when consumers are offered infinite choice, the true shape of demand is revealed. And it turns out to be less hit-centric than we thought. People gravitate towards niches because they satisfy narrow interests better, and in one aspect of our life or another we all have some narrow interest (whether we think of it that way or not)” [3].

At the other end, we have McPhee’s groundbreaking theory of exposure from 1963 [5], in which the author posits the idea that people have different consumption patterns depending on their level of familiarity with the particular field: “the people who choose obscure products tend to be familiar with many alternatives; those who know of few alternatives tend to stick with popular products” [4].

Then where to focus your effort?

In our days of internet, the virality created by instant access to information and to buying products causes narrowing of the head, and leads - at least in the media business - to few mega-hit products.

Actually, according to Anita Elberse’s research which built further on Anderson’s work: “A balanced picture emerges of the impact of online channels on market demand: Hit products remain dominant, even among consumers who venture deep into the tail. Hit products are also liked better than obscure products” [4].

The discussion is of course much more complex and longer than this. Elberse is in fact advocating for a balanced approach, because “finding a good marketing balance between obscure and popular products it’s critical”. It is worth noting that Anderson himself seemed to acknowledge the value and conclusions of Elberse’s research.

So in practice, the trick is to stay active both in the Head and in the Tail regions, and control your product exposure intelligently.

That’s no task for a brain. You need an algorithm. Because you cannot care for your sales without caring for both ends.


Read more

  1. Music Industry Blog, The Death of the Long Tail, 2014
  2. Chris Anderson, The Long Tail, https://www.wired.com/2004/10/tail/ , 2004
  3. Chris Anderson, The Long Tail: Why the Future of Business Is Selling Less of More, Hyperion, 2006
  4. Anita Elberse, Should You Invest in the Long Tail?, Harvard Business Review, 2008
  5. William McPhee, Formal Theories of Mass Behavior, Free Press, 1963
  6. Nicola Salter, The Fashion Industry and the Long Tail Effect, Media Musing, 2013
  7. Robin Lewis, The Long Tail Theory Can Be Reality for Traditional Megabrands, Forbes, 2016
  8. Alfonso Segura, The Fashion Retail Long Tail, 2017