Mobile Advertising

According to Deloitte TMT ( Telecom, Media and Technology ) report, 2016, mobile advertising in India is expected to grace from 2.4 per cent of the overall media expenditure at present to 15-20 per cent by 2020. The growing penetration of smart phones and mobile internet and m-commerce will give a big push to mobile advertising.

The digital population in India has increased from 10 million users to 100 million in a decade but then it has reached to 200 million in just three years. India will have 651 million smart phone users by 2019. In 2011, India’s mobile advertising was $ 25 million. In 2015, mobile advertising is $ 70-80 million. Its annual growth rate is 60-70 per cent. It is the fastest growing segment.

It started with SMS and telemarketing calls. It has now grown into sophisticated in-app ads, mobile web, mobile search and social media. Marshmallon in the latest Android version delivers customised ads on the locked screen of a smart phone. Mobile applications has led to innovations in mobile ads beyond  banner ads. Mobile web and mobile apps work the same way, but the apps are more intuitive, and hence do better targeting. Users spend more time on apps than on the mobile web. India ranks third in downloading apps from GooglePlay. The app budgets of brands are increasing. Many new formats are being  tried, leading to experimentation.

The highest contributor to mobile advertising is going to be in-app ads. Next to it comes mobile videos and mobile ads. Video advertising has been increasing as there is consumption of short format videos on mobiles.

India’s internet access through mobiles is 60 per cent ( IAMAI report.)

Digital Marketing

Digital marketing is through digital channels, which could be online or offline. Online marketing is real time live Internet marketing. Digital marketing could be online, and goes beyond that, say through mobile phones, using apps, tablets and other means of communication such as Podcasts and SMSs. It is a broader term. Digital marketing does not use print media. It uses social media, mobile marketing SEO, QR codes etc. Online marketing is restricted to Internet. Digital marketing includes Internet Marketing, Digital Advertising, and TV, Radio, SMS and Billboards. It has digital foundation.

Internet marketing includes websites, search engine marketing, email marketing, social media marketing, content marketing, email marketing, mobile marketing and banner advertising.

Digital marketing is to be used in conjunction with the traditional marketing.

Digital marketing is more cost effective. Online marketing is measurable. Digital marketing is useful for brand development. Digital  lets us reach the qualified customers. It sets the stage for customer engagement.

Digital advertising is also a channel of digital marketing.

Netflix

On Jan 7, 2016, Netflix has gone live in India. It has a deep and broad library of TV series and movies. It wants to build up on its strength. It has introduced a subscription-based model. The three-tier pricing starts with Rs. 500 a month for a basic pack, Rs.650 a month for a standard pack and Rs. 800 a month for a premium pack. The basic plan allows one person to stream, the standard plan extends this to two persons at the same time, and the premium plan allows four people to stream. The 0.5 megabits per second is the minimum broadband connectivity speed whereas the recommended broadband connection speed is 1.5 megabits per second. This is standard definition and is available in Netflix standard plan. DVD quality recommended speed is 3 megabits per second, and HD quality is given by 5 megabits per second. Over 20 megabits speed is for 4K viewing. The premium pack offers HD and ultra HD content. Netflix  prices are higher than the average cable revenue per user. Netflix basic subscription plan is on par with the premium cable or DTH packs. It is to be combined with high data cost internet plan. The concept of net neutrality could collide with the pricing strategy of Netflix. On the content front, to begin with, Netflix offers the best of Hollywood. It will attempt to make the local content in India. It has an eye on talented people with a strong track record and a passion for storytelling. Acquiring rights to local content may not be easy for Netflix. Most popular content here is owned by broadcasters and producers who themselves have set up OTT platforms.

Machine Learning (ML)

Computers facilitate the activities in life. They till now perform in accordance with the programme and applications. Machine learning breaks new ground. Here the computer performs important tasks by generalising from the available data and examples. It is based on algorithms which can learn from data, with no dependence on rule-based programmes. Machine learning came on the scene in the late 1990s with steady advances in digitisation and economical computing power.

Machine learning is related to artificial intelligence ( AI ). AI means exercising those skills by a computer which requires human intelligence. To illustrate, humans do decision making and visual perception. If computers do it, it is called AI. Computers can be trained to do things which cannot be programmed in advance. Everyday illustrations of machining learning are search engines, Apple’s photo tagging and G-mail’s spam filtering. The basic aim of machine learning is to generalise on the basis of examples in the training set.

Here there are two components of mechanisation in ML. A suitably programmed machine does the classification and prediction. Secondly, the creation of classifier itself is highly mechanised, with the least human input.

AI and ML are not the same. AI is a broad term. Here the computer comes to solve the problems on its own. The information required for solution is coded and AI uses this data to arrive at the solution. ML goes a step beyond. It generalises information from the large data sets and extrapolates and detects patterns to apply the information to new solutions and actions.It is obvious that AI and ML are highly interdependent.

ML enables us to understand the dynamic markets. Organisations do use historical data but they also have to predict behaviour or results. ML gives actionable insights. It makes predictions and finds solutions for business problems — customised recommendations for customers, future performance of employees, forecasting customer loyalty.

ML algos compare new cases to large data bases of similar cases in the past. It is a useful tool for decision makers. Despite this, only humans have critical faculties. ML is made available by cloud providers like Amazon Web Services and Microsoft Azure. These are cloud-based ML services.

Digital Marketing

Digital marketing is through digital channels, which could be online or offline. Online marketing is real time live Internet marketing. Digital marketing could be online, and goes beyond that, say through mobile phones using apps, tablets and  other means of communication such as Podcasts and SMS. It is a broader term. Digital marketing does not use print media. It uses social media, mobile marketing, SEO, QR codes etc. Online marketing is restricted to Internet. Digital marketing includes Internet Marketing,  Digital Advertising, and TV, Radio, SMS and Billboards. It has digital foundation.

Internet marketing includes websites, search engine marketing, social media marketing, content, marketing, email marketing, mobile marketing and banner advertising.

Digital marketing is to be used in conjunction with the traditional marketing.

Digital marketing is more cost effective. Online marketing is measurable. Digital marketing is useful for brand development. Digital lets us reach the qualified customers. It sets the stage for customer engagement.

Digital advertising is also a channel of digital marketing.

Two Types of Programmatic Buying

There are two types of programmatic buying —

  • Programmatic direct
  • Programmatic RTB

Here programmatic refers to automatic. Ad space  is bought automatically on a web page by two methods — bidding for the space or buying it directly.

Programmatic RTB is like Google AdWords. It is only for display ads, and not for search results.

After RTB, the winning ad appears on the web page. Every time an ad loads, it is considered as one impression. Its price is determined by what buyers are willing to pay in real time. This is facilitated by DSPs. A DSP is a fully automated software that bids impressions from an ad exchange.

Programmatic direct is also called programmatic guaranteed and programmatic premium. It is an automated process of buying guaranteed space without involving in an action.

Real-time Bidding (RTB) in Programmatic

By definition, RTB is a method by which advertising inventory is purchased and sold on a per impression basis, through programmatic instant auction, similar to financial markets.

When the bid is won in an auction, the buyer’s ad is immediately put on the publisher’s site. RTB helps to manage and optimise ads from multiple sources ( ad networks ). It provides access to different networks permitting them to create and launch ad campaigns. It prioritizes networks and allocates a portion of unsold inventory, called backfill.

This stands in sharp contrast to static auctions. In dynnmic RTB, the bidding is per impression. In static auction, it is for a group of several thousand impressions RTB is thus more efficient and effective both for the publisher and advertisers.

Imagine a user visiting a site. Here a bid request is initiated. It covers the demographics of the user, his browsing history, his location and  the page being loaded. The request travels from the publisher to an ad exchange, The request and the accompanying data gets submitted to multiple advertisers. They submit their bids in real-time to place their ads. Advertisers bid for each and every ad impression on offer. The impression goes to the highest bidder. The highest bidder’s ad appears on the page the user has visited. The process is repeated for every ad slot available on the page. The whole transaction takes place in milliseconds from the moment the request is received by the exchange.

The bidding process is automatic. The advertisers set budgets for a campaign and the maximum bids. The criteria for bidding for consumers are complex, and consider a lot of data.

Probabilistic models guide us about the probability of a click or conversion, considering the user history. This probability can be used to decide the size of the bid for the respective slot.

DSPs give buyers direct RTB access to multiple sources of inventory. DSPs have the technology to determine the value of an individual impression in real-time — less then 100-milliseconds, considering user history. Large publishers use SSPs to manage advertising yield.

An individual browser history is more difficult to cull on mobile devices. Mobile real time bidding has no universal standards.

The first such RTB platform was launched by DoubleClick Ad Exchange from Google. Microsoft and Yahoo then launched Microsoft Advertising Exchange and Yahoo Ad Exchange respectively.

Data Management Platforms (DMPs) in Programmatic

DSPs use data management platforms or DMPs. They give DSPs a lot of data and information. It is an intelligent technological platform which takes data from first, second and third party of data providers, and manages the data meaningfully so that a DSP can make the best purchase decision on behalf of the marketer.

First party data is owned by the advertisers. It has collected this data from its audience. It is collected from site visitors and existing customers. Tags, pixels, cookies are used for collection . Data comes from o CRM system or web analytics. This data is leverged to  meet the advertising goals.

Second party data is the newest form of data. What is second party data for someone is first party data for some one else. It is made available through arrangements with other organisations of similar target audience. A clothing company can partner with a jewellery company. Here the demographics overlap. It facilitates targeting.

Third party data is data that is not in the realm of first party data, is not connected to the advertiser. An independent party gathers data from diverse web sites, vendors and offline. It is fused with first party data to provide more targeted campaign. Or it can be used independently. It is additional data of consumer behaviour.

The kind of data that is stirred is socio, demographic, behavioral and geographic data.

DMP need to segment the data and normalise the data. It has to find trends. The collection and segmentation of data are together used to send instructions to the DSP. DSP in this light decides the worth  of an impression for the advertiser.

Ad Exchanges in Programmatic

Ad Exchanges carry a pool of ad impressions. These are offered by the publishers in the hope that these will be bought. Buyers then buy the impressions they wish using technology such as DSP. The decisions are made in real-time based on information such as the previous behaviour of the user of an ad being served, time of day, type of device, ad position and many more.

The exchange is a digital market place. It brings together marketers and publishers. It facilitates buying and selling of advertising space; mostly through real-time auctions. They are used generally to sell display, video and mobile inventory.

The examples of exchanges are Right Media ( Yahoo ) and DoubleClick Ad Excange ( DoubleClick ).

Ad Exchanges are a mechanism which make the ad impressions from the publishers available to the DSPs. The price of each impression is decided through real-time auction called real-time bidding ( RTB ). A salesman is not needed to negotiate prices with buyers, since the impressions are auctioned off to the highest bidder.

Demand Side Platforms (DSPs) in Programmatic

This concept has its genesis in Europe. Basically it is a software for purchasing advertising in an automated fashion. DSPs are often used by the agencies and advertisers ( marketers ).These buy display, video, mobile and search ad inventory. DSPs facilitate the real-time bidding ( RTB ) access across multiple sources of inventory. DSPs provide access to more publishers. Previously the process of buying and selling digital ads was manual — there were human buyers and salespeople. They were costly and unreliable. DSPs make the process efficient by automation. It bypasses the need for negotiation of ad rates.

DSPs purchase ad impressions across a range of publisher sites. Still the ads are targeted based on certain attributes — location, previous browsing behaviour. DSPs try to find the best ROAS for an advertiser.