There is interesting news out of MIT. A team used synonyms for words to attack NLP (natural language processing) systems and reduce the quality of results. They tested this on Google’s BERT system to show how their app (TextFooler) can expose areas that need improvement. NLP systems are at the core of Google, Siri, and, Alexa so efforts like this can help them to get better.
The academic paper can be read/downloaded from THIS LINK
“…tiny alterations to the input can flummox an AI and make it misclassify what it sees. TextFooler shows that this style of attack also breaks NLP, the AI behind virtual assistants—such as Siri, Alexa and Google Home—as well as other language classifiers like spam filters and hate-speech detectors” Full post is available at TechnologyReview.com
There have been so many news stories this year about how advertisers can and do target people. Facebook has been making the testimony rounds with the US Senate and EU officials. All this brings to lite the issue of what do platforms know about their users and how do they leverage that information for advertisers so that they can target the right people for their message. As someone that runs many ad campaigns, proper targeting is an important factor when I determine how much budget I am going to devote to my ads. Google has created a new tool for ALL users to be able to go in and SEE how Google is profiling them to advertisers. AND, AND, you have the ability to turn off any of the personalizations that you want to.
You can CLICK HERE or the image below to go see how you are being profiled.
There is a lot of talk in the media these days about creating a paid version of Facebook. Today Facebook makes money from advertisers that would like to target their message to a specific group of people. In an article on MakeUseOf.com they do a good job of laying our how complicated a major change in the unit economics may effect Facebook. Not that my opinion matters to Facebook, but what about a revenue share so that each user gets a small piece of the advertising pie? Facebook would clearly keep the lions share. It is their platform, their servers, their expertise, their ad management programs, etc… but then each of the 2 billion users would share in the advertising revenue. Just a thought.
Today is just another day when a company announces that they have messed up with our personal data. Today is Orbitz that had hackers gain access to 880,000 credit cards from customers spanning from January 2016 to December 2017.
According to the article: “The hackers also likely accessed other data, like names, physical or mailing addresses, birth dates, email addresses, phone numbers and the customer’s gender”
What I don’t see in all these articles is Victim Compensation. When customers fly and the airline has a delay they often compensate the affected people. I have experienced the small ones like when we are waiting to take off and they come around with a free drink or premium snack to bonus miles if the delay is particularly long. When Experian was hacked the compensation was credit monitoring service, not a good compensation for screwing things up, but it is something. Something needs to be done for the victims of these hacks.
Moving to the business side of these issues for a moment. I wonder WHY the public is not informed on what companies like Orbitz are doing to go after the ‘hackers’. I personally would like to know that Orbitz is mad, very mad about this and what they are doing to fight back. The article states; “The company also notes that its current site is not affected by this breach and that it brought in third-party experts and a forensic investigation firm, as well as law enforcement, to “eliminate and prevent unauthorized access to the platform.”” But, there is no mention of going after the data, going after the people that hacked them, it is just trying to stop it from happening again – which I assumed they were doing in the first place.
The old saying in customer service rang true for me recently – It isn’t your fault, but it is your problem. My family has been enjoying the home meal service, HelloFresh for months. We started with Blue Apron. At first we enjoyed the fancy meals and cooking with ingredients that we had never used before. After a few shipments from Blue Apron the kids didn’t love the menu so we decided to give HelloFresh a try. Same concept, but different menu. The first few months were great. The menu was great, food quality was great. We began to depend on our Tuesday delivery and would factor two nights of family meals from HelloFresh into our grocery shopping.
The first problem was two weeks ago. Tuesday came and no delivery. Calling customer service resulted in the response that they don’t know what happened but it would not happen again. The thing is, NO attempt to get the delivery made, just an apology and assurance that it will not happen again. Then, this week the same thing. NO delivery.
I called at 7pm to find out what was wrong (the box had normally been delivered around 1pm-2pm). I was told that the delivery service had until 8pm and their tracking still showed my dinner was out for delivery. 8pm comes no delivery, 9pm no delivery. At 10pm the shipper (a company called DiCom) says that the package was delivered at 2pm. IT WAS NOT delivered.
HelloFresh has a great product, that wasn’t the problem. Their choice of suppliers is what cost them the business.
In current marketing many of the blogs, podcasts and talks at conferences are preaching to “REDUCE FRICTION”. There are several examples of how reducing friction is correlated to increased conversion. In my work it takes a lot of effort and attention to reduce friction wherever possible. The focus needs to remain on your user’s’ experience limiting friction wherever possible and creatively obtain the data that you need without additional burden to your users.
I have been intrigued by the concept of intentionally adding friction into your process. Designing friction into processes and in social media applications, most notably SNAPCHAT. When I was first introduced to SnapChat I found that the company didn’t offer any help, tutorials, or user guides on its website. “The Snapchat app uses a principle Josh Elman from Greylock refers to in his blog post as “shareable design”: the idea that an interface is best understood when explained by someone else.” (source) The ‘shareable design’ seems implement the cognitive bias, Ikea effect. In psychology the Ikea effect is, “The IKEA effect is a cognitive bias in which consumers place a disproportionately high value on products they partially created.” (source). Once an app gets enough users to activate the network effect, using the sharable design to drive user adoption can be a good strategy.
Since the goal of most marketing efforts is to increase conversions, we often focus on reducing friction. When I talk about marketing friction, what I mean is steps and clicks that users must endure in order to do business with your company. If you make an order form across three pages that users have to complete fields and then click to the next page you will loose many people that (while interested in buying from you) are not willing to jump through your hurtles. The site Unbounce has an infographic reporting that forms with 3 fields average 25% conversion, increasing to 5 fields drops conversion to 20% and forms with 6 fields drops to 15% (source). Today Fast Company reported that SnapChat’s CEO Evan Spiegel is “Snapchat is being redesigned to address concerns that the app is “difficult to understand” and “hard to use.”” (source) So while adding friction was used to get SnapChat their initial amazing growth for them to expand their user base they will need to make the app easier for users to use. It will be interesting to see what choices they make in order to appeal to users that are over 35. The next year will be an interesting battle of the social apps; SnapChat, Instagram, Facebook, Twitter, and Linkedin. That is a topic for other posts :-).
It is important to look at your processes from the customers point of view. One thing that I always look for is are we DRIVING or are we GUIDING. It may seem subtle, but it is important to understand which you are doing and which is appropriate for the situation. In customer success there are times when you want to drive the customer to a solution. This is commonly used with support requests. The client would like to have the issue solved, they are unable to solve it on their own and they would like to know that you can take care of them. I do a lot of work training new users on software and in that situation it is important to switch styles to a guide. It is important that the client is an active participant and engaged in the journey.
In your marketing funnel there are also times when you will want to drive and other times where it will be better to guide. Outside the funnel you want to drive the user to your landing page and information. Advertising and social content should drive users at the end into your solution. Once engaged in your solution you want to switch to a guide so that users can see themselves or see their solution in your product or service. Another example is when you are on initial calls or meetings. It is important to drive the start to get to the the heart of their need/pain/problem and then smoothly segue to a guide so that you can uncover the specifics to that clients need/pain/problem and guide them through how your product/service can help them to have a better tomorrow with you.
Last week I attended The Market Research Event in Orlando. The conference has been one of the largest conferences for Market Researchers. After sitting through three days of sessions I was struck by how many sessions that included large brands mentioned the size of their teams and just how few team members they were working with. These are not small companies. For example I attended sessions with speakers from; Dominos, Roche, IBM, Sephora, Lyft, Coke, Colgate, and many others. Each one, at some point in their talks about how they addressed a research challenges mentioned the size of their team. I was surprised at how many global big brands are operating with research teams less than 10 and most with teams of 5 or less.
Another common theme to all the talks was the adherence to a short cycle methodology. Many of the presentations talked about Agile and Lean. Only one from the Garage Group detailed the way they used the build–measure–learn from the Lean methodology but each one had cues to how their internal processes were structured. Going beyond rapid research methodologies and technologies some clients mentioned their usage of communities and platform data analytics.
The talk in the hallways and exhibit halls was about shrinking staff and budgets in the research function. This theme was echoed this morning in Bob Lederer’s daily YouTube report. Bob’s story of the day highlights the major players in market research reporting down numbers. While the numbers may be off a little, they are not off much, many companies are reporting a 5% or less miss of their targets. In my opinion this is a signal to shifting efforts and not a signal to an industry downturn.
Someone on twitter pointed this out for me. The twitter account for Kentucky Fried Chicken has over 1.23 Million followers. BUT @KFC only follows 11 people (6 guys named Herb and 5 Spice Girls) so they have the 11 Herbs and Spices!
I am currently reading “WTF What’s the Future and Why It’s Up to Us” by Tim O’Reilly. Early on I was struck by the passage; “William Gibson famously observed, “The future is already here. It’s just not evenly distributed yet.””. A big part of my conversations with clients and potential clients is uncovering their perspective. I ladder the conversations so that I can understand their ‘today’ and what they want their ‘tomorrow’ to be. It helps me to determine if my company is a good fit to help them achieve their goals. In some conversations they are interested in understanding their data and have only a popular news understanding of machine learning and Ai. In other cases they are very versed in the nuances of machine learning and artificial intelligence and automation as it relates to their data and business. The quote from William Gibson helps me to keep perspective that “The future is already here. It’s just not evenly distributed yet.”