How natural speech tech and content can make or break you, and how to properly assign credit in automotive search engine marketing
While SMX Advanced Seattle Day One was filled with useful SEO, PPC and Social Media nuggets, the second day of the digital marketing conference had a bit of a slower pace, but still plenty of lessons for both the novice and veteran search engine marketer. The biggest takeaways were, again, that natural speech is coming and quality is better than quantity in content. While these two ideas were already set in stone prior to SMX, the conference further reinforced these ideals.
Cortana & The Predictive Search Future
Natural Speech search recognition and optimization is becoming even more important than ever. While Siri was nowhere to be found at SMX, there were no less than two real-time demonstrations of natural speech technology. The first came on Day One when Matt Cutts demoed Google’s voice search function, but on Day 2 when Microsoft demoed the Cortana platform.
What sets Cortana apart is its ability to pull information from all sources on your phone including other apps, schedulers, and more. Cortana, Google Voice Search, and one would assume a next generation of Siri go beyond the usual question-and-answer format of today’s speech recognition platforms. Now, they are, or ill soon be able to predict the information that is required by the user.
Google voice search’s demonstration showed how it could pull information from your original questions into future queries.
i.e. “What is a good Italian restaurant near me?”
“What are its hours?”
“What kind of reviews did it get?”
“How expensive is it?”
Soon, we won’t have to mold our speech around the software as designers are hard at work molding the software around us.
Meanwhile, Cortana displayed similar predictive speech tech. However, what set Cortana apart was the idea that it created a complete profile of you from your schedule, your apps, and even your emails. The idea was that Cortana would be your own virtual personal assistant.
In essence, it could predict if you set the alarm too late for a flight, predict which news stories you were most interested in and more. Of course, there were privacy settings for Cortana as the idea of even a virtual assistant thumbing through your email can be a big turnoff for some potential Windows Phone users.
If a restaurant was used as the Cortana demo and the Space Needle was used for the Google demonstration, then it’s obvious that this tech is most used to find landmarks. That can be easily translatable for the automotive vertical as we need to show the locations of dealerships and garages. Therefore, it is imperative that we properly optimize for these natural speech platforms.
Creating Blockbuster Content
As an SEO, content was of course the theme for SMX. This presentation focused on the idea that too many people within our profession focus on setting specific amounts of content as goals as opposed to quality. This panel argued that Quantity was much less important than Quality in content and named the much-publicized and maligned Facebook updates as an example.
Much has been made of how businesses get little to no organic (or unpaid) exposure on Facebook. The reason may not be necessarily that Facebook is only promoting paid ads by businesses. Instead, there’s simply a mountain of content competing for people’s eyes on Facebook as opposed to years past, so naturally, Impressions are going down.
So, rather than spraying users with a fire hose of decent content, we must instead ensure that each piece of content is excellent, so the eyes that we do get are invested.
Here are some tips:
- Create Long Form Content – Longer content is harder to create, takes more time, and will be better for it
- Have Good Social Signals like Twitter Cards, Open Graph Tags, etc. – Seeing a Facebook Share button or Tweet button can set up the expectation that it’s an excellent piece of content
- Create a source – Another theme of the conference as Author Tags may be becoming ranking signals. If a reader recognizes the author, they are naturally going to be more interested, even if it’s a piece of “thinner” content.
So how do you come up with ideas? Find out from your sales staff which solutions sell and what’s most profitable as that means people are interested in those products. After all, they wouldn’t be buying those products if they weren’t interested.
Another way to get content is to scan through social sites like Google+ and Reddit. Comments on Social media can be a powerful way of teaching you what people are already engaged with in their day-to-day lives. You can apply this strategy as well to competitors. On their Social Channels, what’s attracting negative and positive attention? What isn’t? How can you apply those learnings to your own Social Media properties?
Finally, a great way to create content is to find what common questions are coming in to your call center. Your call center is on the front lines of customer interaction. The best content answers questions, and the questions coming into your call center are the questions that need to be answered.
Executing a Flawless Content Strategy
The main takeaway from this presentation was the idea that not one piece of good content should stand on its own. If it’s an excellent video, it can be cut into a sharable 6 second clip for Vine, turned into a .gif or repurposed another way.
The same goes for written content and infographics. The secret is to repurpose them and “optimize” them for that specific channel. Slideshare for instance is a PowerPoint presentation sharing service and to simply upload an infographic would be a misstep. However, if you could cut that infographic into several slides, well then, you have a winner.
Attribution Success in the Age of Mobile
In automotive, it’s sometimes so difficult to sell digital as a value proposition. How much did that click truly influence the purchasing decision of a $20,000, $50,000 or even $100,000 car? That’s where picking the right attribution model can play an important role in what we, as Marketers, do. That’s why we have to come up with the proper metrics to constantly tell us what is, and is not working in the channels we control.
While this presentation dealt more directly with eCommerce sites where customers buy directly from the site itself, there were several takeaways to be had for the automotive vertical as the ideas can be used for form submissions instead of direct transactions. The entire presentation circled around the idea that the “last-click” model is all but obsolete as it does not tell you the full story of the customer journey through your site. The last-click model itself only tracks the final click where the purchase was made, or, in the case of an automotive site, scheduled an oil change, landed on a car in the inventory, etc.
Instead of the Last Click Model, there are several other models that can be used. One of the more effective methods is the Linear Model since it spreads the credit for the transaction across all of the clicks in the website. However, that too, is somewhat flawed as it cannot completely assign what caused the transaction (for an ecommerce site), or, for an automotive site, a form submission for an oil change, credit application, etc. to take place.
Other potential models included the Time Decay and Position Based models. In the Time Decay model, touchpoints closest to the time of conversion get the most credit. So, while your Social Media channel may have brought customers to the site, it might have been the great deal they saw on the third page of their journey that truly caused the transaction. In the Position Based Model, 40% of the transaction is assigned to the first and last click, i.e. the paid ad and the form submission, while everything in between gets 10%.
Finding which Attribution Model works for you can be a powerful tool in finding the channels that deserve the credit, and in turn, which channels need work.