Prediction - no more long line at cash registers!!

Black Friday shopping is a lot of work. After finding the stuff you want, you have to wait in line. There is a line to try clothes on, to the bathroom, and to pay as well. Why should we still wait in line to pay? I say no more waiting in line to pay!

This idea came to mine as we were waiting in line. We found the exact items online at the company store. We added those to our cart. We were surprised to find bigger discounts online. Hence we just ordered it online when the cashiers refused to match the discounts. Besides the annoyance to having to wait for items, I realized that I should not have to wait in line to pay. I could have self check-out on my phone.

Mobile phones has the technology needed to process transactions. Camera can read the tags on the items. You give it your credit card and your transaction is done and done. I don't need the bottleneck of waiting in line for 30 minutes just to get out of the store. I can be out at other stores spending more hard earned money.

People are comfortable buying on mobile devices. IBM has reported that 25% of commerce on Black Friday is completed online and a lot of it is on mobile. In fact 43% of online came from mobile. So we shouldn't ever need to wait in line to pay again.

Look at Tesco in Korea. In the subway, you scan the items you want on your phone, pay on your phone, and delivery can be done before you get home. US stores need to match the convenience. The more convenient it is for people to buy and pay, the more people want to shop! This last sentence is probably hard to prove, but hey customer is always right!

Inspiration:
1. http://techcrunch.com/2013/11/29/thanksgiving-digest-ibm-e-commerce-edition-mobile-43-of-all-traffic-over-25-of-all-online-sales/
2. http://www.geek.com/mobile/koreas-tesco-reinvents-grocery-shopping-with-qr-code-stores-1396025/

Mobile advertising means native ads, re-targeting options

Mobile advertising is taking off because people spend more time on mobile platforms. The number one activity people do on mobile is arguably social applications. More and more people spend time on social applications on mobile platforms. The trend explains why there is high level of funding activity by venture funds in the space. TapCommerce just raised $10.5 million in Series A round.

TapCommerce aims to increase effectiveness of mobile advertising by targeting and re-targeting customers based on "very large amounts of data coupled with sophisticated statistical analysis." Advertiser wants to know the intent and likelihood of purchase to reach the most likely customers with advertisement. The problem lies in figuring out the intent. Search is the gold standard in customers' expressing intent. Advertising on mobile and social applications have harder time to understand the consumer intents.

The harder to predict nature meant that businesses are turning to analytical tools to dissect correlations and relationships using larger amount of consumer behavior data collected via various commerce site actions.

The reigning king of social network, Facebook, is developing more targeting and engagement methods. Instead of driving just installs, Facebook allows publishers to do call-to-action advertisement as well. User has the highest trust level for ads that appear native. Hence the new expansion allows the action from facebook advertisement to navigate to other native apps for action.

One can argue that switching to mobile advertising is simply chasing where customers' eyeballs. The use of big data analytics is yet to be proven to be more effective as e-commerce platforms collect more and more customer behavior data.


Inspiration:
1. http://techcrunch.com/2013/10/01/facebook-mobile-app-ads-calls-to-action/
2. http://techcrunch.com/2013/11/21/tapcommerce-series-a/
3. http://www.businessinsider.com/native-mobile-ads-dominate-social-media-2013-11

Winner take all in mobile advertising? Or winners by countries?

As people spend more time on mobile devices, advertisers and ad-tech startups increasing want to capture people's eyeballs on those platforms. According to the venture beat article, people on average spend 2 hours per day on their mobile phones. I assume that the statistics is applicable mostly United States. 2 hours per day!! That is more time than I spend on TV! (Note: I only watch TV for about 3 hours a week, and not on TV, but in online streaming sites like ABC.com)

Mobile advertising is a highly competitive space with many entrants aiming to be bigger and better than the previous. All of them aim to help advertiser better monitor ROI and better fine tune the targeting. All of them has incorporated programmatic buy. Programmatic buy just means that you can bid for an ad space in real-time basis based on parameters such as location, time, and user demographics. For example, a gaming app on 7pm wants to show an mobile ad. It sends out a request via the ad network it participates in. The network can potentially shop around different ad networks to find the one with highest eCPM to show. Conversely, programmatic buy tries to find in real-time the lowest CPM while satisfying the targeting parameters for the ad buyer.

What does this mean? Advertising is becoming more and more like the stock market with its many exchanges, high-frequency trading platform, and brokerages. The next stage of development is for exchanges to consolidate to pool together many buyers and sellers. It would provide standardized pricing and bidding strategy for all participants. Think NY Stock Exchange and  Nikkei in Japan. There will be brokerage firms that help ad inventory buyer and sellers to access and find the best price for them in real-time.

Even though the world is flat and getting flatter, there may still many winners in consolidation phase as different parts of the world needing different type of exchanges. Google may dominate the Russian market with AdMob, but TenCent is the powerhouse in China. Let's wait and see!

Inspiration:
http://venturebeat.com/2013/11/11/russian-mobile-ad-investing/
https://support.google.com/admob/v2/answer/3063564?hl=en

Social is mobile, mobile is young

Social happens mostly on mobile. This may be an old adage, but where is the proof? Facebook shelling out top dollars to acquire Instagram; LinkedIn acquiring Pulse. These are all indications that social network titans are working hard to remain relevant on mobile platform. If the users are on mobile, they need to be there to engage with users.

Mobile users are relatively younger as well. You know LinkedIn recently lower age limit from 18 to 13? Grabbing users early on and engage them are the surest way to keep them. If they are on mobile, you need to be there!

However, the Q3 report from facebook painted an accurate data point. 48% or nearly half of daily users of facebooks are now mobile only! The keyword is mobile ONLY. It means advertising opportunity is mobile ONLY.

Advertisement works well on computer browsers where you have lots of real estate to work with. You can show both content and relevant advertisement at the same time. However, this is not true on mobile. If you are facebook, and you see your user base moving to mobile only. The critical thing is to fine-tune the mobile advertising experience on mobile as well. Otherwise, the share of time you have to advertise diminish on computer browsers.

As a consumer, I am more annoyed by advertisement on my mobile device than on computer. On limited screen size, any ads significantly impair my ability to get things done. This is something that companies need to carefully balance out. However, captivated audience is better than no audience. Monetization by mobile advertising will be bumpy for everyone involved.

Inspiration:
http://techcrunch.com/2013/10/30/nearly-half-48-of-daily-users-of-facebook-are-now-mobile-only-says-ceo-zuckerberg/

Beyond cookie to allow advertiser track cross-screen marketing

Mobile-first is no longer the buzz term for advertisers. They realized that mobile like smartphones and tablets are the second and third screens. Consumers use them concurrently as watch shows on TV or computer. Attribution is always the hardest thing for marketers to track. Consumer A sees an ad on TV with call to action; he searches for the interested items on his tablets, but he gives up because it is not optimized for small phone screens; after the show, he goes online on his laptop to search for using different terms. How is this user tracked across different screens?

If every action was indeed based on links in a browser, then cookie and reference-id can be used to maintain identify of the consumer on multiple screen. However, cookie is available only in browsers. TV does not track individual viewers. You can track only by TV-specific call-to-action to track. If consumers do not follow the prescriptive hints, advertisers wouldn't be able to track precisely anymore.

No wonder Google wants to do away with cookie completely. Assuming Google is ambitious as always, I would expect the new system to:
1. Deal with the cookie deletion problem
2. Uniquely identify users on Android platform
3. Tracks users when watching TV like via google chrome cast or google play

This will be a brave new world if Google succeeds!

Inspiration:
http://www.mobilemarketer.com/cms/news/strategy/16455.html

Mobile advertising ROI is better on iOS then Android

Nanigans, one of the biggest buyers of Facebook ads, is reporting a massive gap on advertising ROI. ROI on iOS is 17.9 higher than on Android. The study focuses on retail segments. Let's think about this for a moment.

Assuming their observation is true and is representative of the general populations, I would suggest 2 points:
1. iOS devices commands a premium in the market. Hence its users are more affluent and have more disposable income. However, people with more income not necessarily more inclined to be swayed by marketing messages. My best guess is that the click through rate could be the same or lower in iOS to Android. However, iOS users have way higher conversion since they have more purchasing power to click that checkout button at the end.

2. If this phenomenon of ROI gap persists against more than retailer type ads, then economics forces would change the prices. If iOS is 17.9 times more effective in ROI, the cost would rise over time when compared to Android. The equilibrium point would be that the ROI are the same on both platform. iOS ads cost more than Android to justify for its effectiveness.

Now let's poke holes in Nanigans' observation:
1. Their ad creative do not do well against Android users. We all know you need to test and optimize the creative in your ads to do well. To the extreme, it means the advertiser are not at all effective in writing ads targeting Android users. Probably because people responsible for writing all those ads are using iOS themselves.

2. Retail business is not the same as game and e-commerce industry. Nanigans' data is by no means designed to be representative. It is limited to be retail-focused and draws data only on ads on facebook platform. Could it be that Facebook is simply better at serving ads and targeting ads on iOS platform than on Android?

Only time will tell if any of the above observation is true. However, economic forces will let people exploit the discrepancies in efficiency. If your ads are doing better in iOS, spend more on iOS. If you are better at Android, do more in Android.

Source of inspiration:
http://venturebeat.com/2013/10/16/facebook-ad-profit-a-staggering-1790-more-on-iphone-than-android/

Privacy? No such thing anymore!

This past week revealed trends on where privacy is really dead in this digital age. Google announced its changes to term of use to allow your photo to show up in ads. If you shared about your love for the shoe store in your neighborhood, your picture may start showing up for reviews as people (may be just your friends). On the other hand, facebook announced its decision to drop a feature which allows people to hide themselves when others search for their names.

It is increasingly difficult to remain anonymous online or even just hard to reach. People can say that you can just refuse to use google or facebook. Is it possible? Imagine looking for information without any search engine. Imagine trying to share pictures and comments with your friends via snail mail. The price we pay to enjoy services is getting steeper and steeper. The worst thing is that we do not have a choice. You either accept the term of use or not. There is no negotiation around it.

In a winner-take-all market, the consumers do not have much choice to choose a competitor who would give us different value propositions. Currently there is no choice to get a privacy-enhanced google account or facebook account. It is unclear the disappearing privacy is a good thing or a bad thing in the long run. But the trend is clear.

Inspirations:
Google plus:
http://www.washingtonpost.com/business/technology/google-to-put-user-photos-comments-in-online-ad-endorsements/2013/10/11/322e483e-3289-11e3-8627-c5d7de0a046b_story.html

Facebook change:
http://www.washingtonpost.com/business/technology/facebook-privacy-users-should-check-these-settings-as-new-changes-roll-out/2013/10/11/4a3ef4e2-3274-11e3-89ae-16e186e117d8_story.html

Are you engaged?

The advertising research foundation (ARF) has some very interesting ideas of what they define as engagement in the digital era. Engagement is a new term that sprung up when measuring interaction is now feasible by clicking on a link or expanding a flash display ad.

ARF as “running on a prospect to brand idea enhanced by surrounding context.” This definition is quite abstract. If marketing’s bottom line is to have impact on sales current and future. Then the three fundamental criteria for measurement makes more sense. The engagement metrics has to be reliable, valid, and predictive. The key here is predictive. The measures need to predict future purchase behaviors with good level of confidence.

It is easy to see how the many approaches documented in the ARF papers satisfy the reliable and valid criteria. Some has very good prediction model claimed. For example, Harris Interactive claims the generalized brand health model has 58% correlation with intent to purchase, intent to purchase has 37% correlation with sales. Synovate, another example, can actually use attitudinal equity and market forces to predict purchase probability.

It is always very easy to show correlation when you have lots of data points. However, correlation does not mean causation. Online media and the ease of collecting data means that advertisers have an easier time now to measure the correlation between the metrics they care about with sales. However, there are so many definitions of good metrics that will correlate with sales.

The concept of engagement is the gut feelings by advertisers. Advertisers believe that engagement is the missing secret sauce that would drive future purchase. Engagement is in addition to just brand exposure, attitude towards brand, and so forth. However, advertisers cannot agree on how exactly to measure engagement. More research needs to be done to find the underlying common factor that actually represents engagement and causes future sales. Proving correlation is not enough.

Inspiration:
http://www.thearf.org/research-arf-initiatives-defining-engagement.php

Advertising opportunity on airplane

Finally, we no longer have to power off our devices during take off and landing. A federal advisory committee concluded that most aircrafts today can handle the electronic devices during all phases of the flight.

If passengers can use their electronic devices, it represents even more opportunities to advertise! Think about the captivated audience that will sit in the same place for hours on end waiting to be entertained and reached with great content and even better commercials.

Travelers by airplane represent a more lucrative demographics to advertisers. They are more affluent and have more disposable income to spend. Airlines collected detailed biographical data about the passengers. The segmentation data is really fine-tuned for all the passengers. The general rule of thumb is that the more data you know about the audience, the better targeting you can achieve. Airline passengers disclose gender, age, fare class, travel patterns to the airlines. These kind of data can offer great insights to offer targeted ads.

This is an opportunity area for advertisers to reach the airline travelers better than before. Why can't internet connectivity be offered for free in exchange for opportunities to show advertisement? Customers are receptive to the idea that if a product or service is free, they themselves are the product. Think about free search, online videos. The entire business model of the digital advertising space is to offer free services in exchange for reach to the audience.

The passengers are waiting for good access to the full internet for years now. The technical barriers are coming down. Who is going to step in and capture the full attention of airline travelers? 815 million of them!

Source:
http://allthingsd.com/20130930/aircraft-can-handle-electronic-device-use-panel-to-tell-faa/?refcat=media
http://www.rita.dot.gov/bts/press_releases/bts016_13

Big Data for Real Time Offers

Lily is its name. Lily can apply machine learning to measure attractive score for each offers based on large amount of data gathered for a particular customers. Lily's prowess enables NGDATA to raise $3.3M in its second round of funding. The article explains the mechanism on how the flagship product Lily can offer the best real-time offers to users. From the article, there are two important points worthy of discussion. One is privacy concerns and second is job opportunity, or rather job security.

For real-time offers to work, Lily needs to systematically collect real-time GPS locations from its users. It would also need to know other previously private data such as past credit card transactions, twitter feeds, text messages. These data will help Lily the system to estimate the relevance for an advertisement to be served to a particular customer at a given time.

For example, a customer is walking around central park at 2pm, and it was a hot day. He used his credit card to purchase a nice cold frappuccino from Peet's coffee on a previous hot day. The system would recognize that an advertisement would be relevant if it is from a store with cold refreshments. An example ad might be, "Grab a soft-serve ice cream from ABC store now! 15% off just for you in the next 30 minutes"

Some may find the ads relevant while others may be bothered about the Big Brother surveillance like program to serve ads. If any data is collected and retained by any company, it is only time before government demands it in the name of national security or other purposes. The prevalence of data poses great potential for advertisers, but it also poses great threat to personal liberty and the right to conduct one's life without being monitored. Privacy safeguards are nice, but can never be 100% effective. Only by not collecting such data can privacy breach be 100% eliminated.

Second point is about job security for the consulting services to implement the system. "Channels that Lily can tap up include data contained in subscriber databases and CRM systems; behavioural and transactional data including calls, texts, payments, iDTV, credit card transactions; contextual data such as weather, geo-location, or “inferable context” such as social sentiment or network data; logs from online and mobile applications; third party data such as socio-demographic data; and social media data." All these data are not uniformed and stored in separate storage silos. Integrating these data means consulting service opportunities. Few high level executives and the marketing professionals would even understand the difference between SQL and MapReduce. Using Lily means many many months and even years of data integration work. No wonder data scientists are some of the highest paid professionals among all engineers.

Inspiration comes from:
http://techcrunch.com/2013/09/24/ngdata-series-a/

People spend more time on digital platform than on TV!

From the newsletter "Media Matters" Sept 15, 2013 edition, there is a debate on whether people now spend more time on digital platforms than on TV. eMarketer has made the claim that "“Digital Set to Surpass TV in Time Sent with U.S. Media." If the measures are right and all else being equal, this means that advertising spent on digital platforms should be larger than on TV.

The newsletter criticizes the validity of the average 5:09 hours by guessing what an actual distribution of time spent on digital by percentile. For people who actually go online, it is 7 hours; for top 20 percentile, it is 14 hours per day; for top 7%, it is 20 hours per day. 20 hours per day seem high.

Let us run the same kind estimate for the TV viewing average of 3 hours. For people who actually watch TV, 4.2 hours; for top 20 percentile, it is 8.4 hours; for top 7%, it is 12 hours per day. 12 hours still seem high. Neilson’s people meter only allows 1 TV viewing at once. So how can top 20 percentile people watch more than 8.4 hours of TV everyday? If you sleep for 8 hours everyday, you must be staying home watching TV a lot when awake and not at work.

If we just assume that people spend use 1.5 digital devices concurrently, it will bring the 5:09 hours figures down to 3:33 hours. This revised number is right inline with the TV viewership. If we assume people don’t do anything else, it must suggest that everyone who is watching TV is always surfing online as well. Since Neilson’s 3-hour average already suggests that people do not do anything else in their leisure time watching TV. Hence people must be watching TV and surfing the web at the same time.

Both numbers are unrealistic. The problem with TV and online digital time is that they define the possible maximum. It is only a proxy on guessing if people are paying attention to any media channel. The fact that TV is on or an Internet request goes out and comes back does not represent the fact that people are actually paying attention. Hence, these numbers are all just best effort proxies. They cannot be used to be exact numbers. They can be used to find trends like seeing if people are watching more TV over time or not. However, you cannot compare if people are watching more TV or spending more time online. Both numbers are not accurate measures and they use different assumptions.

Hence, it is definitely false to claim whether people are watching more TV or spending more time online either way. The two numbers are not comparable. This is best solved by conducting a study with following random people around throughout the day and measure their attention on TV or online closely.