Some time ago, I did a blog post called the “Other Long Tail of Search“. When Chris Anderson’s book came out, several smart search marketers realized that the long tail phenomenon applied to our industry as well. If you look at the keywords that drive traffic to your site and plot them according to traffic volume, you get a classic power law distribution curve. A few keywords drive a lot of traffic, but when you add up all the tiny bits of traffic that come from less popular keywords, it adds up to a significant chunk. Suddenly, everyone was talking about the Long Tail of search, and saying why we needed to start paying attention to these dribs and drabs of traffic.
The opportunity of the Long Tail
The premise of my post was that Anderson’s thoughts about long tail markets was not always applicable to search, because of one fundamental issue. Long tail markets are only practical where there’s little or no overhead required to offer the long tail alternatives. MP3’s are a perfect example. It costs iTunes or Rhapsody nothing but a little server space to host an obscure piece of music. And because the purchase mechanism is self serve, there’s virtually no administrative overhead needed. You can offer millions of songs, or billions, for that matter, if they existed, for little more than it would cost to offer hundreds or thousands of songs. All you need is more storage space, and relatively speaking, that’s cheap.
Even if you’re Amazon, and you offer a tangible product, the economics of broadening your selection make the Long Tail possible. In many cases, Amazon, through direct partnerships with publishers or manufacturers, and the opportunities of on demand publishing, can hold the administrative cost of offering a Long Tail selection close to zero. There is no cost until something is ordered, and then it doesn’t matter, because it’s already generated revenue.
And the problem with Long Tails
The search rush to the Long Tail ignored this one fundamental principle. There is significant administrative cost to managing every keyword in a search campaign (we’ll get to organic optimization in a moment). It takes time to set up every keyword in a campaign, and it takes almost as much time to set up the obscure phrases as the popular ones. Broad match options tried to open up the Long Tail, but most sophisticated marketers have found that broad match is not a viable option in most cases. The only place it might be applicable is with branded terms, and even then, by the time you set up your negative match rules, it’s no small investment of time. In most cases, when you consider the traffic the Long Tail phrase will bring in, it’s not worth it.
The same principle guides the choice of search channels you want to utilize for your campaign. Just like with keyphrases, there’s a Long Tail of search engines as well. You have Google at the head, then Yahoo, Microsoft, Ask, and so on, down into the Mivas and that of the world. As we start seeing more vertical search properties emerge, we see more Long Tails. And the curve will change, depending on the intent of the user. For just about every type of search you can name, you’ll generally find Google at the head. But the players that make up the tail can be significantly different, depending on what you’re searching for. The makeup of the Long Tail in the B2B world will be significantly different than that in the consumer electronics world. And if you had to set up individual campaigns on each property, you’d quickly turn in a gibbering shell of a search marketer. Every new search property always comes on board with the same long term revenue strategy. “We’ll use Google’s ad network until we build up enough traffic to launch our own.” The cost of campaign administration is why almost none of them will be able to do it. There’s no shortage of Long Tails in search, but almost none of them lend themselves to the type of economies that Chris Anderson envisions. They’re too much work.
Organic Long Tails
The one place the Long Tail can sometimes represent an opportunity is in organic optimization, but only under certain conditions. If you’re a site that has a huge potential keyword basket (for example, an online store), you have a classic long tail challenge. You may have thousands of products, a small handful of which will drive the majority of the traffic and sales. It’s a no brainer to optimize for these products. But what about the carborundum tipped, left hand curve 6 inch gibbon grooming scissors with the rubberized grip? It may not be a best seller, but if someday someone is looking for one and you’re dedicating warehouse space to it, you want to make the sale. SEO can open up long tail efficiencies if the following are true:
You can make sitewide optimization changes through a CMS solution, avoiding the need for page by page optimization
The vocabulary by which people search for the long tail items is fairly static; and,
When you built out your site, you created the page templates according to SEO best practices.
Only if all these things are true can long tail optimization be worth it. Otherwise, the time cost of optimizing for these products again doesn’t make sense. You don’t get a big enough return on your investment.
In looking at Long Tail curve after Long Tail curve, I suddenly realized something. If you take the click through percentages by position on the search results page (something we gather in almost every study we do at Enquiro) and plot them on a graph, you come up with another Long Tail. The majority of clicks come from the top of the first page, and after that, you get a very long and very skinny tail. Once you move much beyond the first page, the clicks drop to next to zero, but the fact is, when I examine my own visitor logs, I still see the odd person who found my site by clicking on some obscure result from the 5th or 6th page of their Google search.
When I added all this together, I came to the conclusion that because of the large investment of time required to do almost anything in search, Long Tail opportunities were just not that attractive. We’d be better to look at opportunities to move up the tail, or rather, tails. Look for high traffic phrases where you’re lingering at the top of page 2, and concentrate on moving these to page one, and eventually, up into the Golden Triangle real estate. This is true on both the sponsored and organic side. Also, if you’re generating significant traffic on live.com for a keyword, see if there’s an opportunity to capitalize on the same keyword on Google. Always look for opportunities to move up the Long Tail curve, not down it.
Of course, when I blogged about this, it didn’t take long before I got the response (from Cameron) I always seem to get whenever I talk about click through rates in the Golden Triangle spots:
The biggest problem with focusing on position, as you suggest, is that it’s just an attempt to drive up traffic. Focusing on increasing CTR is very myopic. In most cases, CTR has an inverse relationship with ROAS as you’re just getting higher levels of unqualified (or “window shopper”) traffic.
I’m not by any means saying that ROAS is the only key metric (because obviously a very high percent ROAS can happen by sacrificing volume)… but CTR is definitely a false god. The focus has to be on conversion (volume and rate). [emphasis his]
Cameron is absolutely right–you do have to balance conversion and volume. This same point is raised every time I do a panel where we talk about the relative performance of different ad positions. But I believe we’re being equally myopic by not understanding more about why these top spots generate significantly more volume, and why it can be a challenge converting them. I believe, as in most things, that when you take the user’s view, things start to make more sense and opportunities start to present themselves.
Welcome to Search Mall
First of all, consider the search results page like a shopping mall. But this shopping mall has only one entrance. The shopping mall has hundreds of locations, but every time someone walks through the entrance, looking for something, no matter what it is, every story appears to offer it. Some do it much better than others, but at least at first glance, it looks like every story could meet the shopper’s need.
Now, the shopper could spend hours going from store to store, trying to find the best match to their need at the best price. But they’re just not going to invest the time. Again, we have a long tail with significant administrative overhead. And in this case, the curve is determined by convenience and the promise of successfully meeting intent. Convenience is how close to the mall entrance it is, and the promise of success is how quickly the shopper can determine if the store carries what they’re looking for whether the store appears to be a place they’d like to shop at. Attractive stores with relevant signage close to the entrance are going to be at the top of the curve (which shows traffic).
Short listing our alternatives
In making our decisions about which stores we’re going to visit, we’re going to use satisficing, which I talked about in an earlier Just Behave column. We’re going to make a rather arbitrary decision about what’s “good enough.” And the criteria will likely be how far we have to walk to get to the store. The other thing we’re going to do though, and this is critical to understand, is that we’re going to want to compare what 3 to 4 stores have to offer before we buy. When every store appears to offer the same thing, our natural desire for options dictates that we’re not going to buy at the first store we visit.
Let’s add some more texture to this analogy. Right at the entrance, you have 3 stores that have prominent signage saying they offer the lowest prices, guaranteed satisfaction. and the biggest selection. And you have one store (that you can see from the entrance, anyway) that’s much less commercial. In fact, it looks more like a library than a store. And the signage on this location offers objective information about the product you’re going to buy. In fact, the signs say you can talk to other customers who have bought exactly the same product.
Now, with the scene set, imagine yourself walking into the mall. Where are you going to go? Well, you say (quite rightly), “It depends on my intent.” If I were going to the mall to find out more about the product, I would probably go to the library, but I would likely also check out a store or two just to see what the prices looked like. If I were ready to buy, I might skip the library (although the opportunity to talk to other buyers is always compelling) and go straight into the stores. In either case, there’s a pretty good likelihood that the first stores I visit are going to be the ones closest to the entrance.
After I check out 3 or 4 stores, which is a natural behavior, I’m going to make a choice. Did I see an offer in one of the stores that was appealing, or do I have to keep walking through the mall, further from the entrance? The answer to this depends on two things: how diligent a shopper I am, and how good a job the first few stores did convincing me that they had the best match for my needs. For most of us, if there was an offer that looked acceptable, we’d probably call it a day and return to make our purchase. A few of us would continue to search.
But the big variable is how successful the stores were in convincing me that they were my best choice. And again, that success depends on my intent. If I’m browsing and the store hits me with a hard sell, it’s not that great a match. Also, if I’m looking for a iPhone and the store dumps me in the consumer electronics department, with nary an iPhone in sight, I’ll probably beat a hasty retreat and head to the next store.
And there you have a mental framework that allows you to understand the mentality of the typical search engine user. We always start at the top of the page, we always first consider the top 4 or 5 listings, we always at least glance at the top organic result to see if it’s relevant, if these top results are at all relevant, we will always click on at least one of them, and we will always want to consider at least a few options before we make our decision.
It’s still location, location, location
Now, given all that, if you had the choice of location in the mall, where would you choose? Close to the entrance, or buried somewhere in the hundreds of stores far from the entrance.
Now, the analogy needs to be tweaked here and there a bit. First of all, it takes several minutes to adequately check out a store. With a website landing page, you can do it in 15 seconds. This means you can quite efficiently check out more alternatives in a shorter time. But it also means there’s little risk in checking out an alternative. 15 seconds is a relatively small price to pay to check out a website offering.
But perhaps the biggest disconnect in the analogy comes in the area of awareness. If you were a merchant in the mall with a location buried far from the entrance, every day you would walk through the mall, using the same entrance, and walk past the throngs of people jamming the stores near the entrance. You would see them waiting in line to get into the store, apparently eager to buy something. Grumbling, you’d fight your way through the crowds to start your long trek to your store. By the time you got there, the hallways would be almost empty. Sure, you get the odd customer, and a good percentage of them buy (just because yours is the 15th store they’ve visited), but you can’t help comparing your traffic with that of the stores that lie just inside the entrance doors. How long would it be before you’d renegotiate your lease?
But in search, most advertisers have no idea of what they’re missing out on by not being in the top spots. There’s no visual reminder that well over 90% of the traffic is being siphoned off before it ever gets to them. I’m aware of it because I’ve looked at hundreds of eye tracking heat maps that reinforce it. And the traffic that does get to them gets there primarily because the advertisers in the top spot haven’t done their homework and offered the shopper an attractive buying opportunity. They haven’t matched intent.
I remember being at one show where this typical “conversion vs clickthrough” question was raised. Again, the marketer said “Aren’t conversion rates more important?” I said no, it’s the cost per conversion that’s important, which is total cost divided by volume times conversion rate. Let’s take a quick look at each of the factors in this equation, because they’re all equally important.
The cost per conversion formula
First, conversion rates. Will conversions rates go up by moving down the page? Sure, you’ve just forced your customer to run the gauntlet to get to you. Sheer attrition dictates that you’ll have a higher conversion rate. To go back to the mall, if you move your store to the back of the mall, you’ll have a higher conversion rate than the one at the front. You’ll have one hundredth of the traffic, but you’ll have a higher conversion rate. You want an even higher conversion rate? Build a moat around your store and stock it with piranhas. Your conversion rate will soar. Just like you can’t look at clickthrough rates alone, you can’t look at conversion rates alone.
Let’s factor in volume. Even with a lower conversion rate at the top of the page, the sheer increase in volume will create a much higher number of conversions. Yes, it might get frustrating to see a lot of traffic leaving without buying anything, but remember, that’s human nature. We compare alternatives. By moving to the top of the page, at least you’ve become one of those alternatives. It’s far better than never making the short list.
But it’s the final factor that ultimately decides your strategy: cost. Even with a higher number of total conversions, was the increased cost worth it? By moving up the page, your cost per conversion will increase. No doubt about it. But so will the amount of business you end up with. Is the higher lease rate that goes with a better location worth it? That’s a question only you can answer. But I’ll throw out one word of caution here. All too often, when we make decisions about search campaign performance, we compare it against historic search numbers. So, when we move up the page and watch our cost per acquisition rise, we back off. We don’t really consider that the cost per acquisition is still one tenth of what we’re paying through any other channel. What I’m telling you is that there’s potentially a lot of really cost effective business you might be missing out on by not being at the top of the page. Before you decide to pass on that business, take a look at what it’s costing you to buy business through other channels.
Don’t forget that conversion rates can be improved
Finally, we also seem to think that conversion rates are static numbers. They’re definitely not. If you go back to my mall analogy, you’ll see that the biggest factor in success is what happened when a prospect walked in the store. How good a job did the store do in meeting their needs? The better the job, the higher the conversion rate. By moving up the tail, as I suggest, you take your high volume phrases and you optimize not only your position, but also the landing page experience. By focusing on the head, rather than the tail, you’ll get a lot more bang for your search campaign management buck.
So yes, successful search marketing is about balancing volume and conversion, but don’t discount volume and typical search behavior. If anything, this analogy should make you realize that there’s a golden opportunity that sits on top of the search results page. You just have to make sure you don’t blow it by not delivering your prospect what they’re looking for.
Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.