Industrial Strength Search Marketing (SEM, SEO)

So I’ve had this one on the back burner for a while and I guess the time is now…

I’m going just a bit off topic this month. It may seem a stretch, but bear with me, there is a modicum of relevance to this column. You see, most of my ideas for Industrial Strength columns, as well as the bulk of my SEL column writing, take place on either the Capitol 523 or the 542. These are the trains that I take to get to and from work. While I work in Sunnyvale, CA, I live in El Cerrito, just over 50 miles away. Why I live so far from work is yet another long story, so let’s stay on topic here.

All aboard for Silicon Valley

The Capitol Corridor is the name of the rail service I take, and despite the fact that it’s been in operation for 20 years and now enjoys its highest ridership in its history, it always surprises me how few people have actually heard of it. Its more famous Bay Area mass transit cousins are BART and CalTrain. BART is the light rail system that services most of the Bay Area, focusing mainly on customers who commute into San Francisco for work. On a given workday, 300,000 riders take BART. CalTrain, which runs on a normal gauge rail, runs up and down the Peninsula from San Francisco to San Jose and points beyond. It services commuters who live in San Francisco and work in Silicon Valley, as well as those who live on the Peninsula and work in The City. CalTrain serviced 12 million passengers last year.

The Capitol Corridor, which runs on the Union Pacific line from Sacramento to San Jose, carried a mere 1.5 million delighted passengers to work and back during the same period. It is operated by Amtrak, a bit of a throwback, if you will. It’s comfortable, there are table tops, electrical outlets, even a café car. And although it’s not the fastest way to get from A to B, it’s a whole lot better than sitting in traffic on Interstate 880, which is enough to make you want to jam an ice pick in your ear.

I get on the Capitol in the morning and ride down the eastern edge of the San Francisco Bay to Santa Clara. My fellow Yahoos and I then pile into leased vans and drive the 3-4 miles to the main Yahoo! campus. In the evening, we meet at the same vans and head back to the train station. The 542 takes me back up the Bay and drops me a couple miles from my house, where I get into my car and make the short drive home.

Slow going

The major drawback of taking the train, of course, is the speed. It takes me two hours door to door, each way. So if I take the train 5 days a week, every week, that would amount to a whopping 1000 hours of commuting time per year. That’s a lot of time to think about SEL columns! But just like search marketing, train commuting is a process that takes time, and there’s just no getting around it. Regardless of the time and work invested in both, I try to make sure it’s time well spent. As it is on the train, if I mix in a little telecommuting and a minimum of driving (sans ice pick), it’s pretty tolerable. Did I mention they serve cocktails in the Café Car?

Upside: train traveling is a social network

One of the best things about riding the train, other than the obvious fact that I don’t have to hit the road, is the people. My aunt, who grew up in Connecticut, regales me with tales of her father, who commuted to New York City by train for 30 years. From her accounts, the friendships her father formed with his fellow commuters were much stronger than those he had with his co-workers. He changed jobs over the years, but continued to keep in touch with his commuting buddies well after he had retired. I can relate to that.

 This is essentially my social network of follow commuters. I ride with people from a wide spectrum of talents – marketers, engineers, business owners - who by and large work in the tech industry in the Silicon Valley, so it’s a great way to keep up not only with what’s going on in my industry, but also with the tech sector on the whole. Plus, there’s something about sharing your commute to and from work with others that builds deep relationships. I think it’s several factors. First of all, there’s just the amount of time spent together that give you a wide range of topics to ponder and discuss. Second, you never know what’s going to happen once you’re on the rail, and sharing those (sometimes harrowing) experiences deepens bonds between people.

No trespassing: rail blocks ahead

Sometimes people think it’s funny to leave bicycles and shopping carts on the tracks for the train to run over them. While I don’t share that sentiment, these incidents aren’t so bad because if there’s no damage to the train, the conductors can get us going again in 10-20 minutes. The real problem is trespassers. Trespassers are people who end up on the tracks in front of a speeding train.

Most of the time, trespassing is an intentional and final act of a desperate and often depressed individual, but not always. Apparently, it’s quite challenging to estimate the distance and closing speed of a moving train, because some people still try to cross the train tracks in front of us. Some are on their way to school, others play chicken. One unfortunate lady was dragged into the path of an oncoming train several weeks ago by her large dog, which she was walking at the time. Note to self: Don’t walk the dog near the tracks.

As you can imagine, when a trespasser incident occurs, aside from the obvious tragedy (not to mention the trauma caused to the train engineer), the ensuing 2-3 hour delay waiting for the police and coroner to finish their work gives me plenty of time to question the wisdom of rail travel (see Cocktails, above). It’s good to remember that in train commuting, just as in direct marketing, strange things can happen that bring with them unforeseen consequences, and you have to be ready to adjust your strategy on the fly. If it’s not a trespasser on the tracks, it could be a drop in conversion rate or a change in a ranking algorithm that send your results skewing sideways. You just have to roll with the punches and adapt quickly.

How much longer?

Some informal observation tells me that the average shelf life of a train commuter in the Bay Area is about two and a half years (although one guy on my train has been at it over 15 years!). At that point, commuters generally either get another job closer to home, or come to their senses and move closer to work. I’ve been taking the train for three years and change, and although I can’t see myself either changing my job or moving my home in the immediate future, I can’t help but wonder: how long will I last as a rail commuter? I’ve managed to stick with search marketing as long as I have largely because I truly believe that the majority of online marketing will be managed in a similar fashion to Search in the coming years, and that puts me, professionally, in a very good place. So I imagine I’ll be sticking with train commuting, and search marketing, for some time to come.

I’ve been thinking a lot about data-driven SEO these days, there’s so little of it actually going on that I can see. This column was a bit difficult to write, it seems like maybe there should be a book on this, rather than a column. A project for another day….

“How much revenue is SEO driving to our site/property?”

“What’s the ROI on the resources you’re using in SEO?”

“How should we be thinking about funding SEO on an ongoing basis?”

These are the questions I’m hearing these days within the purple walls of Yahoo!

I don’t know about you, but since the economy started heading downhill faster than a snowboarder at the X-Games, I’ve been seeing a lot more attention being given to SEO than in recent years. Back when budgets for SEM were essentially unlimited and the industry was growing like the hat size of a future Hall-of-Famer on steroids, SEO seemed to get the short shrift for a while. I mean, why work really hard on SEO when you can scale SEM so much faster by throwing cash at the problem, especially when SEM budgets are paying for themselves?

Now that the tide has turned, economically speaking, and with SEM budgets focused more on profitability than volume, it’s time to take another long look at SEO and how to use it to your company’s advantage. Of course, in my world, the work starts by Making a Business Case for SEO, where we quantify opportunity with a data-driven approach. Using a conservative valuation of SEO traffic we can not only make the business case for SEO, but also prioritize some of the ensuing projects. I’ve also written about how to take those priorities and implement them throughout a development process, thereby gaining efficiency and lowering cost and time to market. It’s time to take a step back and think about how to manage this on an ongoing basis in a sustainable way.

What I’d like to talk about here is the ongoing tracking of traffic, value, and opportunity required to build out a long-lasting SEO program at a really big company. Because no matter what size company you’re in, here’s what you’re going to face: Once you’ve gotten the green light to proceed with SEO for your site, how do you quantify what you’re actually doing for the bottom line? The truth is that if you don’t know or can’t prove it, your SEO program is about to die a quick and quiet death. The good news is that if you take the time to put into place some decent web analytics that account for SEO, you will have all the data you need to tell everyone in your organization how much profit you’re driving to the company, and where to invest the next dollar.

SEO Referrals

Make sure you have an analytics solution of some kind, almost any kind, on your web property. Ideally, you could have something built just for SEO, but at the very least you should have a package that can isolate SEO referrals from other traffic sources. For us, this meant building a proprietary system that parses SEO referrals from incoming traffic and reports out to users in a bunch of different ways. For most ‘normal’ websites, this usually means something powered by those javascript tags that IT managers hate. They’re easy enough to implement, but from my experience they’re also easy to install improperly and tricky to maintain, and you should know right off the bat that the accuracy of the data they provide is questionable. But that’s OK. Just try to avoid making a data salad out of your results. What I mean by this is try to stick to one source of truth for all your marketing data. Don’t try, for example, to subtract your SEM clicks, as reported by the search engines, from your total search engine referrals number, as reported by your web analytics software, in the hopes of getting a number for organic search referrals. I can tell you this doesn’t work because I’ve tried it, and IT DOESN’T WORK. Instead, choose analytics software that can attempt to decipher whether traffic is coming from paid or organic search and then use those numbers. Similarly, when trying to answer the question ‘how much of my site’s traffic is coming form SEO?’, you will want to use a total traffic number from that same data source, even if you know it is not perfect (see above).

Value

The same analytics software that will segment your organic from your paid search traffic will probably also track your SEO leads to conversion. This of course depends heavily on your business model, but the general idea is that decent analytics can give you a notion of value for your SEO traffic. At Yahoo!, our internal systems track SEO referrals in-session to a variety of revenue events. In some cases the various pieces of this data chain are disjointed, so it’s by no means perfect, but it does allow all properties to track SEO referrals, and in many cases we can track those referrals to direct revenue dollars. I have also written about different valuation models which, when applied to SEO traffic, can help attach real value to organic search referrals.

Reporting

I’ve also written in some depth about quantifying SEO opportunity, in the context of making a business case for SEO. Here I will only add that the same principles apply ongoing to SEO. In fact, once you’ve been tracking SEO data for a year, comparing performance year-over-year will help you make your case for continued SEO work. Since things change so rapidly in SEO, one suggestion I always make in reporting on these ‘long-term’ results is to build out your performance graphs on a timeline and then plot out real events that have taken place – site re-launch, shifts in the market or competitive space, outbreaks of swine flu – it will help you explain otherwise mysterious shifts in your SEO metrics, as well as impress your management.

The advanced track

So once you have the basics covered, you’ll be able to report out on SEO referrals, value, and opportunity. So what’s next? We’re spending some effort trying to understand how the evolving nature of search is affecting our data. For example, of the referrals we’re seeing from organic search, which are coming from web search, image search, video search and so on. We’re calling this vertical search reporting. Also, we’re building out opportunity reports on a fairly regular basis and pushing them out to management periodically as a regular reminder of what’s still on the table as far as SEO value. Finally, we’re trying to combine some of the metrics we have to make more subtle decisions.

Combining Metrics

It’s easy to get mired in all the data, especially at a company like Yahoo!, where the sheer volume of data can be truly overwhelming, but by keeping it simple we can make some useful prioritization decisions.  For example: Search volume is an important component in prioritizing which keywords to go after. You won’t ever be able to get a perfect read on search volume for a given keyword, but there are tools on the market that can give you a general or relative idea of search volume. Rank, when paired with search volume, can help you prioritize keywords and their untapped potential. You’ll need to make some assumptions about clickthrough rates associated with each position. By combining these metrics we can prioritize keyword level optimization efforts. This helps us understand for a given keyword, how much effort we should spend optimizing our site, given the value of the opportunity for that word.

Keep it simple

First things first, remember the basics:

1. Get an analytics solution in place that can isolate SEO referrals
2. Attach value to SEO referrals through tracking technologies or valuation models, or both
3. Find useful ways to leverage the above data to build useful reports for your management
4. Rinse and Repeat