Skip to content

Digital Marketing II: Inbound, Social and Mobile

Innovation & New Business Planning - TUHH Institute of Innovation Marketing & Institute of Entrepreneurship, Hamburg · part of my Technology Management MBA · study notes for revision.


The previous chapter was about buying attention. This one is about earning it. The reading’s own image for the difference is the cleanest summary anyone has written: outbound marketing is like hunting for a needle in a haystack, while inbound marketing is like holding up a magnet so the needle comes to you. Instead of chasing potential customers with advertising, the firm positions itself as the thing people are already searching for.

Two forces push in this direction, and both are measurable. The first is that advertising is losing its grip: roughly 86 percent of US television viewers skip the commercials completely, and 84 percent of consumers aged 25 to 34 have clicked away from a website because the ads on it were irrelevant or intrusive. The second is that search has taken over the start of the buying process: about 82 percent of US consumers do online research before they buy anything. If people are already looking, the cheapest thing you can be is findable.

This chapter runs through the rest of the digital picture: getting found, creating content, converting on the landing page, then the five separate jobs that make up social media work, then what genuinely changes on a mobile screen, and finally the two supplemental themes that sit underneath all of it - big data and data science, and the uncomfortable trade between personalisation and privacy.

1 · Inbound marketing: the shape of the idea

Section titled “1 · Inbound marketing: the shape of the idea”

Outbound marketing fights for the best ad placement on the results page. Inbound marketing fights for something else entirely: making sure that when someone types a keyword, the firm’s organic links rank higher than competing organic links. Same page, different half of it, completely different economics.

Outbound covered in the previous chapter
  • The company starts the conversation and pushes a message out
  • Competes for the paid slot on the results page
  • You pay per click or per thousand impressions, every time
  • Reach scales with spend, so the cost never goes away
Inbound this chapter
  • The customer starts the conversation by searching
  • Competes for the organic rank, which cannot be bought directly
  • You pay to create content and to earn links, mostly once
  • Built on blogs, podcasts, white papers, video and search engine optimisation
The reading defines inbound marketing as engaging consumers by creating content so that the company, its brand, its products and its services are found when consumers go looking for information.

The section then organises itself around three questions a marketing manager has to answer, and the next three sections are exactly those answers.

How do consumers find me?What content should I create to attract them?How do I optimise my website for engagement?

2 · Getting found: how a search engine decides who wins

Section titled “2 · Getting found: how a search engine decides who wins”

Start with the brutal fact that makes this worth any effort at all: organic links appear in a specific order, and research shows most consumers never look beyond the top few organic links on the first page. Rank three is a business. Rank thirty is invisible. The process of climbing that list is called search engine optimisation, and the reading defines it plainly as using features that make a company website rank higher in the organic listing without paying any money.

To play the game you have to know what the engine actually does. It does two basic things.

Indexit catalogues the web pages it finds, exactly like a library catalogue
→
Rankit orders those pages by relevance and importance for the query typed
→
Resultthe ordered list of organic links the consumer actually sees
Indexing decides whether you exist at all in the engine’s world. Ranking decides whether anyone will ever see you.

Ranking rests on two key factors, and the split between them is the single most useful thing in the section, because the two factors are earned in completely different ways.

Relevance on your own page
  • How closely the engine can match a consumer’s query to one particular page
  • Driven by the page title
  • Driven by the page content
  • Driven by meta tags, the lines of code in the head section of a page that carry extra information about its structure
  • You control all of this yourself, which is why it is the first thing to fix
Authority on everybody else’s pages
  • The engine’s measure of a page’s importance
  • Based on the number and the authority of other web pages that link to it
  • Measured by the proprietary PageRank algorithm, which grew out of a doctoral thesis by one of the search engine’s cofounders
  • The idea is borrowed from academic citation: a paper is important if many papers cite it, and a citation from a heavily cited paper counts for more
  • You cannot set this yourself, you can only earn it
The two ranking factors. One lives on your own page and can be edited today, the other lives on everybody else’s pages and has to be earned.

So inbound marketing needs two key actions, one per factor. First, make sure the website carries the content, keywords and meta tags that raise its relevance to a specific query, which means anticipating the keywords and search phrasings consumers are likely to use. Second, find ways to earn inbound links from other websites in order to build authority, and the way to do that is smart content creation. Notice that the second action hands the ball straight to the next section: authority is not an SEO trick, it is a by-product of publishing something worth linking to.

3 · Creating content: every firm is now a publisher

Section titled “3 · Creating content: every firm is now a publisher”

Content creation for digital marketing has two main goals, and they map onto the two ranking factors almost one to one.

Goal one: answer the customer’s questionsgo beyond basic product information and solve the consumer’s actual problem
↓
Goal two: be original, trustworthy and interestingso that other organisations and individuals choose to link to you, which builds authority
Goal one feeds relevance and the customer relationship. Goal two feeds authority. Content marketing is doing both with the same material.

Goal one asks for a customer-centric view of the world. You look for ways to solve a consumer’s problem whether or not it connects directly and immediately to a sale. The reading makes the point with unusual honesty: even if that consumer eventually buys from a competitor, you have still improved your relationship with them and still raised awareness of the product and of the need for it.

  • Home Depot, the world’s largest home improvement specialty retailer, thought about how people actually approach a home improvement project. To mirror the practical advice its staff give in the stores, it built videos and blogs on its website giving general home improvement advice. As the site’s relevance to consumers rose, so did its traffic.
  • Cosmetics companies moved from simply advertising products to publishing online video that teaches people how to apply makeup. The video answers a genuine informational need and introduces the products naturally, without sounding like a commercial script.

Goal two asks the content to build brand reputation, because reputation is what makes other sites link to you.

  • Whole Foods blogs not only about groceries and recipes but about organic and sustainable living.
  • American Express runs a platform of content on broad business topics such as leadership, marketing and digital tools, with the ambition of becoming the go-to expert not just on credit but on every aspect of running a business.

The conclusion the reading draws is worth memorising as a sentence: every firm has the potential to become a publisher. And if you are a publisher, you need what a magazine has always needed, namely a designated editor in chief who owns the creating, managing and distributing of the online content. That is an organisational decision, not a marketing tactic.

The formats named across the section are the practical menu: blogs, podcasts, white papers, videos and the search engine optimisation work that makes all of them findable.

4 · Optimising landing pages: the other half of the battle

Section titled “4 · Optimising landing pages: the other half of the battle”

Getting the consumer to click the organic link is, in the reading’s phrase, only half the battle. The other half is deciding what appears on the landing page so the visitor converts into a customer. At that point user experience is paramount.

The recommended method is to test the design with consumer focus groups, and the reading lists exactly what is worth testing.

ColoursButton shapesPositionWordingFontsImagesLogosUsabilityPerformance

The payoff can be large and it is not only commercial. When a US government agency redesigned its website for usability, the task success rate almost doubled, the time to completion fell, and user satisfaction soared. And because smartphone use keeps rising, companies are shifting to a mobile-first mentality for website design, which is the bridge into section 11 of this chapter.

A/B testing is the tool the reading names for this kind of work. It defines an A/B test as an experiment in which the impact of a single variable is tested, commonly used on ad copy and on landing page copy or design, to find out which version drives the desired result better. The discipline is in the word single: change the headline and the button colour and the image at once and you learn nothing about which of them mattered.

5 · Social media: why word of mouth became the main event

Section titled “5 · Social media: why word of mouth became the main event”

Outbound and inbound are both about company-to-consumer interaction. Social media is different in kind: it is a tool that lets people create and share information and ideas across the globe, mostly with each other, and the company is only one participant.

The reading proves the scale with events rather than statistics. Facebook and Twitter were instrumental in the Arab Spring uprising that began in December 2010, and Facebook played a significant role in the 2016 US presidential elections and became a focal point of the investigation into Russian interference. In public health, the Ice Bucket Challenge for amyotrophic lateral sclerosis research started in Boston and went worldwide: between 1 June and 1 September 2014, more than 17 million challenge videos were shared on Facebook and were viewed over 10 billion times by more than 440 million people. In entertainment, the 2012 pop video Gangnam Style passed two billion YouTube views within two years and had reached 3.29 billion by February 2019, the sixth most-watched video on the platform. Before the May 2019 release of the film Detective Pikachu, an actor tweeted about a supposed leak of the full movie on a consumer YouTube account, and the studio’s marketing team, which presumably owned that account, posted a clip of Pikachu dancing for the film’s full running time of 1 hour and 44 minutes. Within two months it had been watched almost 25 million times.

For companies the consequence is captured in a much-quoted line: a brand is no longer what we tell the consumer it is, it is what consumers tell each other it is. The numbers behind that sentence are the ones to remember.

What was measuredThe finding
Share of purchases driven primarily by word of mouthAn estimated 20 to 50 percent of all purchases
Peer effect on app adoptionA person is three times more likely to adopt a mobile app if a friend adopts it
Conversations and brand sales, study of 170 brands, 2017Online and offline conversations drive an estimated 19 percent of brand sales, equal to between 7 and 10 trillion dollars of annual US consumer spending
Value of a word-of-mouth customerCustomers acquired through word of mouth are worth twice as much as those from other channels
Peer effect on churnProbability of defecting to a competitor rises by 80 percent if a customer’s friend has defected

Social media has also produced a new class of social influencers and digital celebrities: a beauty expert with 38 million Instagram followers by mid-2019, and an internet personality and actor with 21 million Instagram followers and 35 million YouTube video views. Companies increasingly use them to support brands, and in some cases these celebrities have gone on to launch brands of their own.

The section then splits the management job into five distinct tasks, and the next five sections take them one at a time.

Listenwhat are they already saying
→
Participatejoin and shape it
→
Amplifyget the message spread
→
Measurework out if any of it did anything
→
Manageprevent and survive the bad days
The five social media jobs. They are usually done by different people, and only the last one gets attention when it fails.

Conversations on social media contain information companies can use, and the reading sorts that value into three uses.

Understanding customers. Surveys use predefined questions and focus groups sample only a handful of people. Social media gives a broader and deeper view of consumer attitudes and behaviour, because it is what people say unprompted. The example is PatientsLikeMe, a platform where patients going through significant treatments such as organ transplants, or living with chronic conditions such as ALS or multiple sclerosis, share personal experience and hear from others with the same condition. Patients get medical advice from their doctors, but they are hungry to learn and share about self-care issues such as managing side effects or pain. Novartis partnered with the platform in a significant way to build a community of organ transplant patients and learn from their experiences. Such partnerships let companies detect and analyse consumer needs and then tailor products, pricing and communications. By 2019 the platform had become the world’s largest personalised health network, with more than 650,000 people covering 2,900 medical conditions and generating more than 43 million data points, which is why UnitedHealth Group acquired it in June 2019.

Brand perceptions. How do consumers see the brand, what do they like or dislike, what do they associate it with? The traditional answer is a custom study producing a perceptual map. But millions of casual social conversations are also a window into the consumer mind, and using sentiment analysis and text mining a company can build brand association maps in real time. The reading shows one for Nike.

Customer satisfaction and future behaviour. Reviews and conversations both reveal satisfaction and predict future purchasing. Consumers routinely check hotel reviews on TripAdvisor or restaurant reviews on Yelp before booking, and the effects are measurable.

StudyResult
Yelp restaurant ratingsA one-star increase was associated with a 5 to 9 percent increase in sales
Television showsShows discussed across a wide number of communities did better than shows discussed inside a single focused community
One online communityInteractions between members raised consumer spending on books, music and movies, especially among the most active and connected participants
Multichannel entertainment retailerA double-digit increase in spending was attributable to the purchase behaviour of consumers who joined the firm’s online community

Listening is passive. A company can also join in and shape the conversation.

  • Starbucks built a product suggestion site, My Starbucks Idea, to source new ideas from customers. It connects enthusiasts with each other and lets them take part in the brand community, while employees review the content and implement selected ideas.
  • C Space runs hundreds of online communities, recruiting members who match the target profile of its corporate clients, and actively facilitating the discussion so those clients learn about customer behaviour and unmet needs. The snack manufacturer Nabisco launched its 100 Calorie Pack after exactly that kind of input.

Participation also opens the door to real-time marketing. When a blackout halted the 2013 Super Bowl, quick companies turned the moment into reach. Walgreens tweeted that it does carry candles. Oreo posted that the power being out was no problem because you can still dunk in the dark. In five hours that day, Oreo’s Instagram following went from 5,000 to 34,000.

8 · Leveraging and amplifying: why, what and how people share

Section titled “8 · Leveraging and amplifying: why, what and how people share”

Companies can use social media to amplify a message and create brand advocates. Tremor, a company started by P&G in 2002, uses a community of more than one million mothers, along with boomers, millennials and US Hispanic consumers, to generate word-of-mouth recognition and advocacy from influential consumers. Most campaigns now combine paid and social media.

Two campaigns carry the numbers worth quoting.

CampaignWhat was doneResult
New Zealand Police, 2017A force that was 80 percent white and 70 percent male wanted more diverse recruits, so it made a YouTube video featuring a diverse group of 70 of its officers and promoted it through influencersReached over 91 million people within a week, top trending video on YouTube and Facebook, web traffic to the recruitment site up 898 percent, applications from diverse populations up 615 percent, and eventually the largest number of female recruits ever
Wendy’s, 2018A mixtape of rap tracks aimed at McDonald’s and Burger King, promoted on Twitter and released on Apple Music, Google Play and SpotifyClimbed to the top 50 of Spotify’s Global Viral list and to third on the Apple Music hip-hop charts, and collected 76 years’ worth of streaming in one week

To amplify anything deliberately, a manager has to understand three things: why consumers share, what they share, and how sharing spreads.

Why. Motivations are intrinsic or extrinsic. Intrinsic motivations are intangible rewards such as social status: Yelp runs special events for elite members who contribute reviews, and that VIP status motivates them to advocate for the site among their friends. Extrinsic motivations are tangible incentives such as money or free products, which is exactly why influencers with large followings, who can earn substantial income, work hard to be first with new products and trends.

What. To trigger the intrinsic motivation, content has to connect emotionally. Research on viral video ads found a repeatable recipe.

Open with a joyous tonegrab the viewer instantly
↓
Build an emotional roller coasterhold attention through the middle
↓
Surprise, but do not shockshocking videos are interesting, yet people are less likely to share them
↓
Target people likely to sharethe reach depends on who receives it first
The viral-ad structure named in the reading. Humour usually works: an Amazon commercial showing celebrities testing improbable uses for Alexa collected almost 40 million views in seven months.

There is a trap here that the reading is careful about. Highly entertaining ads are shared more often, but that is not automatically good for the company, because the brand message can get lost in the entertainment. Brands have to balance the entertainment that earns the share against the branding that earns the association and the purchase intention.

How. Can you engineer virality? There is a real argument about this. One social media agency guarantees that it can engineer virality. The evidence points the other way.

EvidenceFinding
2012 study of millions of messages on Twitter and YahooMore than 90 percent did not spread beyond the recipient at all; about 4 percent were shared exactly once; less than 1 percent were shared more than seven times
YouTubeThe average video gets fewer than 10,000 views; as of 2017 there were almost 11,000 videos with over a billion views, a tiny fraction of the billions on the site
Diffusion rateVirality borrows from epidemiology, where each infected person infects more than one other. Most studies find the diffusion rate of social media content is far below one

That is the distinction between virality and amplification. Buzzfeed’s founder argues that social media messages can be amplified two to three times, but they cannot really go viral. If the multiplier is two or three rather than exponential, the strategy changes completely: the size of the initial seed matters more than the hope of a chain reaction.

This is the argument against paying for a handful of star influencers. A 2011 study tracked 74 million events from 1.6 million Twitter users and concluded that, under a wide range of plausible assumptions, the most cost-effective performance comes from ordinary influencers, that is people with average or even below-average influence. The recommendation is a big seed strategy: seed the message with a large number of average users instead of a few celebrities.

The debate has not stopped brands from paying anyway, and the prices are worth knowing.

Influencer post, 2016, 3 to 7 million followers
up to 187,000 dollars per post on YouTube
Same spend expressed as cost per thousand impressions
CPM of 26 to 62 dollars
YouTube average for comparison
CPM of 7.60 dollars

So an influencer post costs somewhere between roughly three and eight times the platform average per thousand impressions. Many marketing managers pay it anyway, on the belief that influencers generate more engagement than other paid media. Success has also let social celebrities launch their own brands, including the beauty subscription service Ipsy, formerly MyGlam, plus Kylie Cosmetics, Huda Beauty and the Italian shoe company Attico.

9 · Measuring social media: why the obvious numbers lie

Section titled “9 · Measuring social media: why the obvious numbers lie”

It is hard to measure the impact of social media on sales, so most firms fall back on surrogates: the number of followers, fans, shares and likes. Those counts are often converted into earned media, meaning the amount of money the firm would have had to spend to expose an equivalent number of people to the message.

The reading then takes one famous number apart. In 2013 a social analytics company reported that the average value of a Facebook fan across several brands was 174 dollars. The method was to compare the spending of people who had liked a brand with the spending of people who had not, and to call the difference the value of a fan. Two problems destroy that inference.

Self-selection the direction of causation
  • Heavy drinkers of a cola are more likely to click Like on that cola’s page in the first place
  • So liking may not cause extra purchasing at all
  • The causality may run entirely the other way: buying a lot causes the like
Homophily birds of a feather flock together
  • People with similar preferences end up in the same community or group
  • Your friend buys the album two days after you: was that your influence, or simply that you both already had the same taste?
  • A study using instant messaging data from 27 million users found homophily explains more than 50 percent of what looks like contagion
What happens when you control for both
  • Lab and field experiments on over 14,000 users controlled for self-selection and homophily
  • The conclusion was that the dollar value of a Facebook like is close to zero
The honest position
  • Because of these challenges, estimating social media effectiveness is still an open research area
  • Follower and like counts are surrogates, and should be reported as surrogates rather than as returns

10 · Managing social media: dry forests and fast responses

Section titled “10 · Managing social media: dry forests and fast responses”

Social media is a double-edged sword. Consumers can promote a brand and spread your content, generating awareness and goodwill. They can equally broadcast a bad experience, and when negative word of mouth strikes a chord in a network it spreads like wildfire and becomes a public relations disaster. Even if 99 percent of customers are happy, one customer can create a nightmare.

The best advice, the reading says, comes from the people who fight forest fires, and it is two rules.

Rule one: do not let the forest get dry prevention
  • You never know where lightning will strike, so the only defence is that there is nothing dry to burn
  • If one complaint catches the attention of thousands or millions, that is the signal of a dry forest, meaning a genuine underlying problem with the product or the service
  • We live in a transparent world and companies cannot hide from consumers
  • Avoiding the disaster before it happens is a management responsibility, not a communications one
Rule two: if it catches fire, move fast response
  • 42 percent of customers expect a reply to a social media complaint within one hour
  • 80 percent expect a reply within 24 hours
  • Almost 50 percent say they would stop doing business with a firm that does not respond
  • Complaints negatively influence 62 percent of customers
Prevention is a product and service question. Response is a staffing question, because a one-hour expectation cannot be met by a team that reads the mentions once a day.

Two cases show both halves. After a musician’s guitar was damaged by mishandling on a United Airlines flight, he wrote a song about it that became a YouTube hit in 2009 and generated serious bad press. The effect was strong enough that the airline now uses the video in its own employee training programme, which is the dry-forest lesson learned the expensive way. In 2013 a Taco Bell employee posted a photograph of himself licking 30 taco shells to his Facebook page and the national news media picked it up. The company responded quickly: it stood behind its food-handling policy, fired the employee, and verified that the shells in question were never served to customers. That is rule two executed properly.

Organisationally, two things follow from all of this. Content needs an editor in chief who owns creation, management and distribution, as section 3 said. And the reading closes the essential part by naming organisational structure, alongside consumer behaviour, engagement and measurement, as one of the things that has to be rethought creatively for digital.

11 · Mobile technology: what actually changes on a small screen

Section titled “11 · Mobile technology: what actually changes on a small screen”

Mobile use has grown dramatically worldwide since the iPhone launched in 2007, and in emerging markets such as India and Indonesia the majority of people met the internet for the first time on a phone rather than a computer. Whole businesses now exist only because of mobile - Waze, Uber, Postmates, Grubhub, HotelTonight - and established companies are rebuilding their sites with a mobile-first mentality. The question the reading asks is whether mobile is simply a smaller screen, or whether it changes behaviour. The answer is five changes.

1. Screen sizehow much people will read
Small screens make people less inclined to read for long, so mobile suits visually rich content: photos, maps, videos and gameswhich is why Facebook leans on photos and video rather than text in the newsfeed and on Instagram
2. Locationthe phone knows where you are
Phones supply location data that sharpens targeting: a consumer within one mile, about 1.6 km, of a store is 23 percent more likely to click that store’s adalmost every major company now uses location data to target
3. Price transparencythe shop is no longer a closed room
Shoppers can check competing prices and reviews while standing in a physical store, which produced showroomingexperience the product in the shop, check the price on Amazon, buy from Amazon; US retailers have taken real bottom-line damage from this
4. Apps, not browsingthe biggest strategic break
90 percent of smartphone time goes to apps rather than the web, so display and search ads reach a smartphone user only 10 percent of the timethe desktop playbook simply does not transfer, and mobile strategy has to shift
5. The phone is personalso ads read as intrusions
By December 2016 there were 600 million devices running ad-blocking software worldwide, 62 percent of them mobile, with the blocking rate up 30 percent in a yearconsumers are getting better at avoiding advertising no matter how good the story is

Two responses to that last point are named. Mastercard’s chief marketing officer argues for a shift from storytelling to story making with consumers, meaning the consumer takes part rather than watches. Google argues for moment-based marketing: wait for the right moment, then send the right message. It classifies these micro moments into four intents.

I want to knowI want to goI want to doI want to buy

The examples are concrete: book a ride on a Friday evening and restaurant or cinema ads become relevant at that instant; driving on a motorway at midday is the right moment for a map app to show a nearby restaurant. Red Roof Inn used the same logic to turn flight cancellations into hotel bookings.

The apps problem, and the mobile metric problem. Global mobile advertising spend was expected to grow from 231 billion dollars in 2019 to 396 billion dollars in 2023, and yet the mobile playbook cannot be copied from desktop. In-app and banner ads are often felt as intrusive and irrelevant. Worse, the standard yardstick misleads: measuring mobile ads by click-through rate can be wrong because of the fat finger phenomenon, the unintended or inaccurate taps that touchscreens produce. And getting an app onto a phone is hard, because a typical smartphone user carries about 40 to 50 apps and uses only 10 to 15 of them regularly. Nobody has any reason to download an app for a cola or a bar of soap.

The essential reading closes with a warning that is easy to skip and important to keep: the digital landscape moves so fast that best practices even a year old can be obsolete, so the durable skills are creative thinking about underlying consumer behaviour, about reaching and engaging consumers, about measurement, and about organisational structure.

The supplemental reading defines both terms crisply. Big data means large volumes of varied data arriving at high velocity, the so-called 3Vs. Data science is the analytic process of exploring and modelling that information to inform marketing decisions. Two examples already appeared earlier in the reading: analysing browsing behaviour to adapt a website to a visitor’s cognitive style, and using social media to build a picture of multiscreen content consumption.

There are two challenges before any of the benefits arrive.

Challenge one: the right data and the right technique
  • No matter how big it is, the wrong data skews the answer. A large survey of Xbox users produced an atypical prediction of votes in a US presidential election; once demographics were controlled for, the bias disappeared and the prediction matched other polling
  • Big data cannot account for all consumer behaviour
  • Some problems need data visualisation and exploratory analysis, others need statistical or machine-learning models
  • Some questions can be settled by correlation, others need a causal model or an experiment
  • Choosing the right tool needs expertise, and that expertise is in short supply
Challenge two: it is an organisational problem
  • IT has to receive the data and warehouse them properly
  • HR has to hire the talent that can extract insight, and build a structure where findings actually circulate
  • Marketing and operations have to turn the insight into revenue or into cost savings
  • This is why many firms have a chief information officer who reports to, or works closely with, the chief marketing officer
Neither challenge is about the size of the data set. One is a methods problem and the other is an org-chart problem, which is why big data projects usually fail for non-technical reasons.

Once both are solved the payoffs are large, and the examples are grouped by function.

FunctionExample and result
PricingMacy’s uses real-time demand information to price 73 million online products hourly
PricingFICO is exploring alternative data such as social media to price credit risk, especially in developing countries where traditional credit-scoring inputs do not exist
PricingUber uses surge pricing that adjusts dynamically to geography-based supply and demand
PricingAirbnb built a predictive model using property and neighbourhood information to suggest rental prices to hosts
PromotionPantene varied its targeting by postal zip code weather patterns, matching the product advertised to humid or dry conditions, and sales rose 28 percent
OperationsUPS optimises driver routes, saving 100 million driving miles a year, cutting fuel by 10 million gallons and carbon dioxide by about 100,000 metric tons
OperationsGeneral Electric sensors on jet engines and wind turbines could produce savings in the hundreds of billions of dollars across several industries
TalentKnack.it offers consumers games whose in-play choices generate data for behavioural matching with employers, and it accurately predicted which employees would produce the best innovation ideas at one company

The conclusion is a management one rather than a technical one: big data and data science can transform decision making, but only if several business units cooperate to apply the right analysis to the right data.

Personalisation demonstrably works, and the reading gives three pieces of evidence.

Matching a website to the visitor’s cognitive style
purchase intent up by almost 20 percent
Ads that address the individual directly
almost double the CTR of ads targeted only on demographics and interest
Retargeted ads shown at the right purchase stage
higher CTR than the same ads shown at the wrong stage

Targeting for direct marketing is an old tool. What digital adds is two things: it incorporates far more detailed data, and it automates the personalisation at machine level, on the server or on the smartphone itself. That combination puts a great deal of power in a marketer’s hands, and that is exactly where the privacy problem starts.

The consumer position is genuinely split. 61 percent of American consumers and 48 percent of UK consumers say they are willing to share personal data in return for personalised marketing communications, and in general people are comfortable sharing with brands and companies they trust. At the same time they are increasingly worried about their information being compromised, periodic data thefts damage even the trusted brands, and a lack of transparency plus frequently updated or reversed privacy policies feeds the doubt. On the regulatory side, the reading notes only that some countries, especially in Europe, impose stricter rules than the United States on collecting and using private information; it does not name a specific law.

There is one more twist. Consumers are starting to ask why third-party vendors and market research firms should profit from selling data about them without sharing any of it back. A new group of companies, such as Streamr and UBDI, offer platforms where people can sell their own personal data. One startup cofounder estimates that an individual’s data might be worth almost 2,000 dollars a year, although other estimates are much lower.

The venture. A small hardware startup selling a compact indoor air-quality sensor to schools and small offices, with a subscription dashboard. It has almost no budget, so inbound has to carry the top of the funnel.

The content plan. Four pieces, each aimed at a specific search intent rather than at the product, each feeding one landing page, each with a single conversion metric.

Content pieceSearch intent it targetsLanding page it feedsConversion metric
Guide: what CO2 levels in a classroom actually meanKnow: someone typing a question about safe carbon dioxide levels indoorsClassroom monitoring page with a two-field demo requestDemo requests per 100 visitors
Checklist: how to ventilate an office without heating the streetDo: a facilities manager trying to solve a problem this weekDownloadable checklist behind an email fieldEmail captures per 100 visitors
Comparison: the four ways to measure indoor air qualityBuy: someone in evaluation mode comparing approachesProduct specification page with pricingAdd-to-cart or quote requests per 100 visitors
Case write-up: one school, one term, the measured resultBuy and trust: a buyer who needs a reference before committingBooking page for a fifteen-minute callCalls booked per 100 visitors

Notice that only the last two mention the product at all. The first two exist to satisfy the reading’s first content goal, answering a real question regardless of an immediate sale, and to earn the links that build authority for the pages that do sell.

Organic versus paid, side by side. Take a target of visitors to the site, and compare buying them with earning them. These figures are my own illustration, not from the reading.

Paid searchOrganic content
Up-front build cost02,400 (four pieces at 600)
Monthly cost1,200 (1,000 clicks at 1.20)0 after the build
Visitors in year 112,000about 3,300, ramping from zero to roughly 700 a month
Cost per visitor, year 11.20about 0.73
Visitors in year 2 with zero further spend0about 8,400 if the ranking merely holds
Cost per visitor across both years1.20about 0.21
Paid traffic
visitors = monthly spend / CPC, and it goes to zero the month you stop paying
Organic traffic
cost per visitor = one-off build cost / cumulative visitors, which falls every month the page keeps ranking

That is the whole argument for inbound in one table. Paid is rented reach: predictable, instant, and it disappears the moment the card is declined. Organic is an asset: slow, uncertain at the start, and it keeps paying afterwards. The sensible venture runs paid to learn which keywords and messages convert, then spends that knowledge on content aimed at the same intents.

  1. Write down the three inbound questions for your own venture and refuse to move on until each has a real answer: how do my customers find me today, what content would genuinely help them, and what does my landing page have to do to convert them.

  2. List the actual search phrases your buyer would type. Not your product category, the words a person uses when they still have a problem and no solution in mind. This is the anticipation step that drives relevance.

  3. Fix relevance first, because you control it. Page titles, page content and meta tags on every page that matters. It is cheap, fast, and it is the only ranking factor you can change unilaterally.

  4. Plan authority as a slow campaign, not a task. Decide what you could publish that another organisation would voluntarily link to, and remember the reading’s standard: original, trustworthy, interesting, and contributing to your brand reputation.

  5. Name your editor in chief. If you are going to behave like a publisher, one person owns creating, managing and distributing the content, or it will not happen.

  6. Build one landing page per intent and A/B test one variable at a time. Test wording, button, image, colour, layout and speed, but change a single thing per test or you will not know what worked.

  7. Listen before you post. Find where your customers already discuss the problem, run sentiment and text mining if you can, and use it to check your assumed brand associations and unmet needs.

  8. Choose one participation move you can sustain. An idea platform, a facilitated community, or simply answering fast and well in public. Sustainable beats clever.

  9. Plan amplification as a big seed, not a lottery ticket. Assume the multiplier is two to three, not exponential, and put the budget into reaching many ordinary sharers rather than one expensive celebrity, unless you can show the engagement justifies the CPM premium.

  10. Set your measurement honestly. Report followers, shares and likes as surrogates, state that self-selection and homophily inflate them, and pick one behavioural outcome such as demo requests that you would defend in front of an investor.

  11. Do a dry-forest audit and a response plan. What is the real product or service weakness that could catch fire, and who replies within an hour, in what tone, with what authority to fix things.

  12. Test everything on a phone before a desktop. Assume visually rich content, assume attention is short, work out whether you need an app at all, and if you do, name the unique value that justifies the download.

  13. Write your data promise before you collect the data. What you collect, what you use it for, what you never do, and where the personalisation stops feeling helpful and starts feeling intrusive.

TermWhat it means in plain words
Inbound marketingMaking the firm findable so customers come to it, rather than pushing ads at them; the magnet instead of the needle in the haystack
Search engine optimisationThe work of making a website rank higher in the organic listing without paying the search engine anything
IndexingThe engine cataloguing pages on the web, like a library catalogue, so they can be retrieved later
RelevanceHow closely the engine can match a query to one page, based on page title, page content and meta tags
AuthorityThe engine’s measure of a page’s importance, based on how many other pages link to it and how important those pages are
PageRankThe proprietary algorithm that scores authority, built on the academic idea that a citation from a heavily cited source counts for more
Meta tagsLines of code in the head section of a page that carry extra information about the page’s structure
Landing pageThe page a clicked link delivers the visitor to, whose only job is converting that visitor into a customer
A/B testingAn experiment that changes exactly one variable, usually ad copy or landing page copy or design, to see which version performs better
Earned mediaThe reach you get when customers, the press and the public share your content or talk about you, valued as the money you would have had to spend to reach the same number of people
HomophilyPeople with similar preferences cluster together, so what looks like one friend influencing another may just be shared taste
Self-selectionThe people who like or follow a brand were already its heaviest buyers, so the comparison flatters the channel
Big seed strategyAmplifying by seeding the message with a large number of ordinary users rather than a few expensive influencers
Amplification versus viralityReal social content typically multiplies only two to three times; true viral spread, where each sharer produces more than one new sharer, is rare
ShowroomingExamining a product in a physical store, checking the price on a phone, then buying it online elsewhere
Micro momentsGoogle’s four intent-rich moments when someone reaches for a device: I want to know, I want to go, I want to do, I want to buy
Fat finger phenomenonUnintended or inaccurate taps on touchscreens, which make click-through rate an unreliable mobile metric
Big data and data scienceLarge volumes of varied data arriving at high velocity, the 3Vs, and the analytic process of exploring and modelling it to inform decisions
  1. Explain what inbound marketing is, name the two forces the reading says are making it more important, and give the supporting numbers for each.
  2. What are the two things a search engine does, and what are the two factors ranking is based on? For each factor, say what you can do about it.
  3. Give the two goals of content creation, and say which ranking factor each one serves. Name two examples from the reading.
  4. What is the difference between virality and amplification, and what does that difference imply for how a small venture should spend its social budget?
  5. A consultant tells you a Facebook fan is worth 174 dollars. Give the two reasons the reading says that number cannot be trusted, and say what a controlled study found instead.
  6. Name the five ways mobile differs from desktop, with one number for each where the reading gives one.

Next: Pricing: The Basics → - what a price actually has to do.