This got me thinking about marketing an unproven product.  Setting expectations too high could cause a huge backlash that could ruin a company or product.  Nowadays with twitter and ratings sites like yelp, a set of unhappy customers all responding to good marketing is hard to bounce back from.  Any tips to manage those expectations from the start without losing market share?  What about ways of fixing a fiasco?

Share/Save

As a marketer, I read David Mullen’s 5 PR Predictions and think he has some great points.  I want to focus on his point about measuring people.  I agree with David that we will start measuring people instead of impressions, but I think it will go a lot further than that.  I don’t think the economy is turning around so quickly, which will cause companies to be very “bottom line centric”.  Companies will want to show how their investment in Web 2.0 has paid off and that will mean qualifying the active users/visitors.  It will mean developing ways to report how effective social networking tools are, like managing a twitter account.  Yes, twitter allows the company to talk to it’s customers one-to-one, but did that person buy something?  Did they click on an ad?  Did they refer someone?  How long did the cycle take? 

I think 2009 is going to be the year of the number crunchers and web analysts.  Maybe it is time to take a statistics refresher course.

Share/Save

This time, instead of looking at the existing portfolios, I looked at the individual keyphrases to see how they performed and if there would be any obvious answers.  Luckily there was!  It turns out that over the last 4 months more than 95% of the 22,000 keyphrases didn’t result in a sale, an assist to a sale or lead.  This made my head spin.  I had always assumed that during the sales cycle people were clicking on the paid ads multiple times.  Therefore, even though we only expected a few hundred sales a month we would need a few thousand active keyphrases to support the multiple clicks of a single user.  This assumption was proven wrong by the numbers.   

After uncovering this flaw in our strategy, we have now restructured the campaigns into only 4 groups:  Performers, Assisters, Lead Generators and Growth Phrases.  The performers can be directly tied to a sale.  The assisters can only be connected to a performer, meaning they were clicked on early in the cycle.  The lead generators only caused leads and the growth phrases are phrases we hope will become part of one of the first 3 groups.  This allows us to put 95% of budget into performers and assisters, 2-3% for lead generators and another 2-3% for growth phrases.   This restructuring allowed us to cut 23% of the budget with room in the budget to experiment on new phrases, even during tough times.  It still isn’t clear how much of a dip in revenue we will see.  My hope is that there will be less than a 10% drop in revenue, because the only group that is truly cut is the the outliers, an unpredictable bunch. 

I think this strategy can work for all companies in times like this.  You may need different labels on your groups and maybe more than 4 groups, but the overall strategy should be the same.  Each group should have its own budget and target ROI associated with it, letting you optimize each one as frequently as possible.  We are planning on regrouping every month based on 4 months worth of data.  These two parameters are probably different for different companies, because it all depends on the sales cycle.  Companies with a longer sales cycle and a higher end product can probably afford restructuring monthly.  Shorter cycles may demand weekly restructuring.

I am really excited about this update and if anyone decides to implement something similar, please let me know how it works for you.  If you already have, let me know if there is something I am missing or if I completely missed the mark, please let me know too.  Feedback is always welcome!

Share/Save

Yesterday’s NY Times article wrote about the talks and mentioned that the biggest thing holding things up are the privacy concerns that Verizon has.  The authors mentioned that Verizon hasn’t decided how much data about their customers they want to give over to Google.  I am happy Verizon is thinking about my privacy, but I am still concerned.  What data does Google want?  What kind of data is Verizon thinking about giving over?  Is it only about searches?  Does it include my subscriber status?   

I am going to hope that Verizon is only thinking about giving over information related to searches, because even the thought of more than that scares me.  More than search related data includes: how much I pay a month in usage charges, how long I use the phone for a day, how much money I spend in transactions over my phone, how long I have been a customer, if my GPS is on then where I travel, what my local hangouts are and anything else they know about me.  

In a comment to my post about where privacy and marketers meet, Roman said that I shouldn’t fear marketers because the data they have or want is relatively harmless.  I agree, I do not fear marketers.  I fear ad providers and ad networks because their networks and ad space is worthless if they cannot serve ads that will attract the right customer’s attention.  In order to optimize their ad server technology, they would love to know customer demographics, buying power and search and click patterns.  

If the deal goes through will Verizon allow their customers to opt-out of all the data sharing it is doing with Google?  I am sure not, because Google wants to know where you are when you search. They want to be able to serve you ads for a store right in front of you, or a restaurant down the block or reenforce a banner that is on the side of a building across from you.  Not only will they have the ability to serve local ads to you, they will also be able to interpret the meaning of your searches based on proximity.  For example, a search for the word “Prada” will be more valuable if you are on Fifth avenue in New York City than if you are in the Fort Lauderdale airport.  

At the end of the day, I hope we figure out how to curb all this data sharing and tracking across different media and that Verizon Wireless acts responsibly.

Share/Save

This made me realize that I am a little Jekyll and a little Hyde.  The Dr. Jekyll in me thinks that no one should be tracked if they don’t want to be and everyone should be able to choose how they are marketed to.  The Mr. Hyde in me loves the idea of tracking users, segmenting customers and using machine learning to enhance marketing campaigns.  This seems like a clear contradiction, am I a schizophrenic? In hopes of convincing everyone reading this that I am not a hypocrite, I will try to rationalize my thoughts.  

I believe that the constitution provides everyone with an inherent right to privacy, including online privacy, and should be forced to actively opt-in to a service where you give up that right.  This means that a privacy policy on a website should not automatically opt you in to a  tracking service.  On the other hand, I also believe that marketing strategists should be able to tailor their campaigns to as small a customer segment as they can.  This means that if the strategist selling jewelry knows who shopped for high-end clothing recently and only wants to display an ad to that specific customer, they should be able to.  The only caveat to my second statement is that everyone who is shown those customized ads has to opt-in to them.  

Most marketers don’t agree with me because it would make their lives harder.  Not only would they have to ask before being allowed to track a customer, they would also have to ask before showing a customized ad.  I think most marketers assume that if people are forced to make a conscience choice, they avoid making it and use the default settings.  This would cause marketers to have a lot less data about their customers and would not be able to make as educated a guess when choosing where to place an ad.  

I, on the other hand, believe that people make conscience choices about everything they do and would welcome another one.  I also believe that having less tracking data to work from may mean more work for the marketer, but it will also result in a better product and a better ad.  This is where scientific marketing comes to play.  Marketers should start by making a hypothesis based on a small amount of data and then try to prove it with experimentation and watching more data.  In effect, scientific marketing is a little behavioral targeting and a little marketing psychology, but it conforms to a world where privacy matters!

Share/Save

Now that I am opted-out of seeing the ad, what is stopping Yahoo! from taking my search and surfing patterns and then creating a model to market to someone that is not opted out?  If there are lots of people like me going from Toyota’s site to the WSJ, what is stopping Yahoo! from going directly to Toyota and letting them know?  This way, Toyota will just buy more ads on WSJ.  Maybe it will cost more, but chances are it will be more effective advertising than placing the ads randomly.  

Despite my negative twist, I would still like to commend Yahoo! for being the first provider to allow such an opt-out service!  Customized advertising costs more for advertisers and while Yahoo! is lagging behind Google in ad sales, this is a brave step.  I think Google should take a page out of Yahoo!’s book and allow similar opt-out services on their sites, otherwise I think they are risking being labeled “evil”.

Share/Save

For the climate issues, I think we should have the people who created the Truth campaigns do it.  They have been successfully riding a very fine line with their campaigns.  They are all witty, educational and make people aware of the hazards of smoking cigarettes, marijuana and doing drugs.  

On the other hand, for AI I would love the Axe or even the Mac marketers working on our cause.  They could make it sexy and cool and then all the kids would want to be working on it!

Share/Save

I think Al Gore’s work to make people aware of global warming has been a fantastic marketing campaign.  His movie, ‘An Inconvenient Truth‘, started to raise awareness and the public conversation about global warming and climate change.  In the last few months, he helped launch an organization called We Can Solve It, whose motto is “we can solve the climate crisis”.  This organization is marketing heavily and has not only made people aware of the issues, but is also instructing people how to help.  This means, they are trying to rally a grass roots following to solve the problem.  I think they have the right idea.

Imagine marketing AI as a problem that we haven’t solved yet, but can be solved!  Imagine a marketing effort that not only educated the people, but also wanted to activate the masses.  I think that in this day and age, one where web 2.0 and collaborative spaces have proven successful and extremely useful, we should be able to do it!  Now all I need are a few people to help me get the ball rolling.

Share/Save

I agree with him that the media has taken some fantastic accomplishments and focused on the wrong or easiest aspects, but how could they have known which was the hardest or easiest task?  In the article he refers to a picture of the Sony Aibo dancing after having played a successful game of robot soccer autonomously.  Doug points out that the dancing was the easiest part of the entire task to do, which he is right about.  Does that matter to most people?   I don’t think most people realize that making a robot dance is less complicated than making it find a ball and kick it.  If you were to think of it in human terms, most people would classify both tasks as similarly complex.  Dancing involves moving with some sort of coordination and so does kicking a ball.

I don’t think the media is to blame, I  think it is the AI communities fault for not marketing their accomplishments properly.  We try to get each other’s respect and recognition, but never look to the rest of the world.  Most people don’t know how hard it is to accomplish true AI, they only know what they see in the movies and in the papers.  I work as a consultant and when I mention that I used to work with Elizabeth Sklar, PhD developing educational robots, most people start talking about the Roomba.  This is their association with robotics.  The AI community has already done bigger and better things, like the DARPA challenge, but hasn’t marketed it enough.  We see AI and Machine Learning in use in our everyday lives, like on recommendation engines on Amazon.com and Netflix.com, but since they aren’t highlighted and marketed as AI tools, no one notices them or refers to them when AI is discussed.  

I think that if we market our AI accomplishments and make them easy for everyone to use or contribute to, the community will gain tremendously.  I think a marketing effort could take a generation of web-savvy children and make them into AI researchers or developers.

Share/Save