Internet Marketing Integrating Online and Offline Strategies

1. 1Diagnose where you sit on the brand-strength line before you design anything. The source states that the stronger the brand, the easier it is to get a prospect to accept an offer, and that an unknown brand finds action harder to compel. Rate your brand as established or unknown, because that decides how much trust you must build inside the offer itself.
2. 2Separate the two time horizons of trust. Long term, trust is earned by how you do business: prompt and accurate fulfillment, satisfying returns and service practices, and guardianship of the customer's privacy. Short term, the marketer without an established brand needs a way to make the prospect feel reasonably comfortable about transacting now. Build one plan for each horizon.
3. 3Name the specific fear that blocks the transaction. Amazon's early customers were reluctant to hand over credit card numbers on the web, so the fear was financial loss from transmitting card data. Write your own blocker in one sentence (card data, privacy, delivery, returns) before choosing a remedy.
4. 4Choose the risk-reduction tool that matches the named fear. The passages list a strong guarantee, customer reviews, third-party endorsements, and a bigger incentive. Pair a guarantee with a financial fear, as Amazon did when it promised to reimburse any losses from transmitting card information. Use reviews and endorsements when the fear is whether others were satisfied.
5. 5Make the guarantee prominent, not buried. Amazon posted its guarantee prominently. A reassurance the prospect never sees reduces no risk.
6. 6Use trust-borrowing channels with disclosure. Third-party blogger communications are often viewed as highly trustworthy, even though ethical bloggers reveal payments and gifts, and in the United States they are required to. Treat disclosure as part of the trust asset, not as a cost to it.
7. 7Back the promise with operations. Review fulfillment speed and accuracy, the returns experience, and privacy handling. These are what the source says trust ultimately depends on, so a guarantee on top of weak operations is a liability.
8. 8Re-evaluate the short-term device as the brand matures. Amazon's card-loss guarantee suited its time and was later replaced by a free shipping offer on many purchases. Schedule a review so each device still matches the current trust gap.

Amazon in its early days. The company was making Herculean efforts to establish its brand, yet many customers still hesitated to give their credit card numbers over the web. Amazon named the fear, loss from transmitting card data, and answered it with a prominent guarantee to reimburse any losses customers incurred from sending their card information. That gave a prospect with no experience of the brand a reason to feel reasonably comfortable transacting. The source calls the move appropriate for its time. As the brand strengthened and the fear faded, the guarantee was replaced by a free shipping offer for many purchases. The long-term foundation remained the same: prompt and accurate fulfillment, satisfying returns and service, and guardianship of customer privacy.

• If your brand is unknown, then add explicit risk reduction (a strong guarantee, reviews, third-party endorsements) to the offer, because a weak brand makes action harder to compel.
• If the prospect's hesitation is about transmitting payment data, then guarantee reimbursement of losses and display it prominently, as Amazon did.
• If you use bloggers or other third parties to carry your message, then require disclosure of payments and gifts, because trustworthiness depends on it and U.S. rules require it.
• If a short-term trust device (such as a guarantee) has done its job and the brand is established, then shift to a simpler incentive such as free shipping, as Amazon did.
• If you are tempted to rely on assurances alone, then check fulfillment, returns, service and privacy practices first, because long-term trust comes from how you do business.

• Treating trust as a brand-awareness problem only. The tell: heavy advertising spend while fulfillment errors, slow shipping or poor returns persist.
• Offering a bigger incentive when the real barrier is perceived risk. The tell: discounts increase but conversion at the payment step stays flat.
• Hiding or omitting the guarantee or reassurance. The tell: the policy exists but sits in fine print and customers still voice card-security worries.
• Using influencer or blogger content without disclosure. The tell: payments or gifts are not stated, which erodes the credibility you paid for and breaks U.S. requirements.
• Keeping a dated trust device after the context changes. The tell: you still run an old fix for a fear customers no longer hold, while a current friction (such as shipping cost) goes unaddressed.

• ▸Brand strength: Rate as established or unknown, and state the evidence.
• ▸Primary customer fear: Write the one concern that blocks the transaction (e.g., card data loss, privacy, delivery, returns).
• ▸Short-term device: Pick a guarantee, reviews, third-party endorsement or larger incentive that answers that fear.
• ▸Placement: State exactly where the device is prominently displayed in the customer journey.
• ▸Fulfillment standard: Define what prompt and accurate fulfillment means for you and who owns it.
• ▸Returns and service practice: Describe how returns and service are handled so customers are satisfied.
• ▸Privacy guardianship: List how customer data is protected and how this is communicated.
• ▸Third-party voices: List bloggers or reviewers used and how payments or gifts are disclosed.
• ▸Review date: Set when to check whether the short-term device is still needed or should be replaced.

1. 1Map every touchpoint in the customer lifecycle where evaluation forms — order placement, delivery, service inquiry, self-service — because satisfaction is cumulative across touchpoints, not a single moment.
2. 2Diagnose the operational cause of dissatisfaction before touching messaging: at Herman Miller the root was lack of synchronization between eight plants, not poor communication. Trace long lead times, haphazard delivery, and errors to their process source.
3. 3Install the systems that give customers real-time visibility and control — FedEx let customers schedule and track packages in real time, and that degree of control directly increased satisfaction while cutting routine call-center inquiries.
4. 4Set a hard operational rule tied to the satisfaction driver: Herman Miller synchronized every item to the longest production date and forbade building any item more than three days early, then measured on-time shipments as the proxy for satisfaction.
5. 5Share your data with the customer (mirroring): Frito-Lay electronically shares store-level inventory and pricing data with retailers so both parties hold identical data, which reduces procurement costs and errors.
6. 6Use point-of-purchase data to tailor the experience per customer — Frito-Lay built individual planograms per store, cutting SKUs 20–30% while increasing convenience and satisfaction.
7. 7Aim the whole CRM effort at a seamless customer experience across channels (the Disney multichannel example), then track satisfaction metrics like on-time rates, error rates, and complaint volume as leading indicators.
8. 8Listen to complaints and reverse decisions that hurt experience — Frito-Lay pulled its noisy compostable SunChips bags in 2010 after customer complaints, choosing satisfaction over the sustainability feature.

In 1995 Herman Miller had long lead times, haphazard delivery, and poor customer service, producing a measurable decline in customer satisfaction while profits fell nearly 90% in a single year. The root cause was a lack of synchronization across eight plants, creating a three-to-four-week gap between when the first and last components of an order were completed and driving 8-to-11-week internal lead times. The company installed planning software with the explicit goal of improving satisfaction through faster service, querying each plant for a firm completion date and synchronizing every item to the longest date with a no-build-more-than-three-days-early rule. Within 18 months on-time shipments reached 99% and inventory turns doubled. Five years later its Internet-based division installed orders in three days to two weeks with a 99% on-time record and virtually no errors, and inventory turned 40 times a year versus an industry average under 20. Satisfaction here was rebuilt not by marketing but by fixing the operational synchronization behind the touchpoints customers actually experienced.

• If dissatisfaction is measurable, then trace it to a process/synchronization failure before rewriting messaging — Herman Miller's problem was coordination, not communication.
• If a touchpoint leaves customers uncertain about status, then give them real-time visibility and control — FedEx tracking raised satisfaction and cut call-center cost simultaneously.
• If a product feature generates complaints, then weigh it against overall experience and be willing to reverse it — Frito-Lay pulled the compostable SunChips bags.
• If you and the customer work from different data, then mirror your data to theirs — Frito-Lay's shared store-level data cut procurement errors.
• If you have point-of-purchase data, then personalize the offer per customer/store rather than serving a uniform assortment — individual planograms raised both convenience and store sales.

• Treating satisfaction as a marketing overlay while operations stay broken — the tell is high complaint volume alongside good ad performance (Herman Miller's award-winning reputation coexisted with collapsing satisfaction).
• Optimizing one plant or channel in isolation so components arrive out of sync — the tell is finished goods waiting weeks in warehouses for the rest of an order.
• Withholding status information from customers to protect the call center — the tell is high routine-inquiry volume that a self-service, real-time view would eliminate (FedEx cut this).
• Adding a feature customers dislike and defending it on principle — the tell is complaints spiking after a product change (noisy SunChips packaging).
• Underrating self-service and channel cost — the tell is rising service cost per contact when cheaper channels could raise satisfaction; the passages flag cost of customer service by channel and anticipatory service as themes.

• ▸Touchpoint: name one point where the customer evaluates you (order, delivery, service, self-service).
• ▸Root operational cause: the process/synchronization issue behind current dissatisfaction.
• ▸Satisfaction proxy metric: the measurable indicator (e.g., on-time %, error rate, complaint count).
• ▸Baseline vs. target: current value and the number you commit to (e.g., 99% on-time).
• ▸Visibility/control given to customer: what real-time data or self-service you provide.
• ▸Data mirroring: what data you share so you and the customer hold identical information.
• ▸Personalization lever: how POP or CRM data tailors the experience per customer.
• ▸Complaint-reversal trigger: the complaint threshold at which you pull or change a feature.
• ▸Channel cost/fit: cost per contact and whether the channel raises satisfaction.

1. 1Define the acquisition event before choosing channels. The definition counts two outcomes: a first purchase, or a behavior that permits further contact, such as an email opt-in. Write down which one you are buying. If the second, name the exact permission level you are collecting (the source treats permission marketing as having levels).
2. 2Profile the target with behavioral data first, demographics second. The source calls demographics 'necessary but not sufficient' and behavioral clickstream data 'key'. Pull demographic data (e.g., Pew Internet & American Life for U.S. users) to learn who is online. Then use your own clickstream data to see what they do from information search through purchase.
3. 3Settle the motive question separately. Clickstream shows what people do, not why. The source says motives are 'only available from marketing research', so ask a sample of the target before you build messaging.
4. 4Check reach before spending. The source notes that older, poorer, less educated and rural segments are less likely to be online. If a large share of your target audience is offline, assign it to other media rather than forcing it through digital channels.
5. 5Build the channel set from the tools the source lists for acquisition: search (SEO and PPC), display advertising with ad targeting, vertical ad networks, social network advertising, email, affiliate programs, portal relationships, event-driven marketing and viral marketing. Start with search and email. The source says most consumers start the purchasing process with search, and over 90 percent of Internet users use email and search daily.
6. 6Pair every acquisition channel with a destination built for the event in step 1. For leads, that means a tested landing page and a defined conversion path, with the persona and purchasing scenario written first. Ensure the page works on the devices your target uses, since the source says marketers must reach customers 'on the device they prefer'.
7. 7Treat the acquisition front end as a test program. Direct-response practice in the source separates the front end (acquiring the customer) from the back end (what the customer later does). Test offers, creative and targeting, and track outcomes by channel.
8. 8Judge cost by customer lifetime value (CLV), not by first-sale margin. The source's direct-response chapter places CLV at the center of how much acquisition spending is justified. Compare each channel's cost per acquired customer or opt-in to the CLV of customers it brings in.

The source gives no figures for a complete acquisition campaign, so this is an illustrative scenario built from its framework; the numbers are assumptions. A merchant wants new customers. Step 1: it defines the event as an email opt-in, with a first purchase to follow in the back end. Step 2: clickstream shows its buyers start with search and check email daily, which matches the source's finding that over 90 percent of Internet users do both daily. Step 3: it runs paid search (PPC) and optimizes organic content, sending both to one landing page built around a single opt-in offer. Step 4: it adds a small test of social network advertising and an affiliate partner, each with its own tracking. Step 5: after a set period it compares cost per opt-in by channel, then the share of each channel's opt-ins that purchase. Step 6: it prices each channel against CLV, not first-order profit. A channel that looks expensive per opt-in but sends buyers who return stays in the plan. A cheap channel whose opt-ins never buy is cut.

• If the target's behavioral data shows search as the entry point to purchase, fund search (SEO and PPC) first and treat other channels as additions.
• If the acquisition event is a permission-based contact, then design the opt-in around a stated permission level and respect CAN-SPAM requirements before sending anything.
• If a meaningful share of the target is offline (older, rural, lower-income), then do not expect digital channels to reach them and budget other media for that segment.
• If two channels have similar cost per acquisition, then rank them by CLV of the customers each delivers, not by front-end volume.
• If you can describe who the target is but not why they buy, then commission marketing research on motives before scaling spend, because clickstream data cannot supply it.

• Segmenting on demographics alone. The tell: you can describe your audience by age and income but cannot say what they did before purchase.
• Counting clicks or impressions as acquisition. The tell: reports show traffic growing while first purchases or opt-ins stay flat.
• Sending paid traffic to a generic page. The tell: strong click-through with weak conversion, and no landing page or conversion path tied to the offer.
• Judging channels on first-sale cost and ignoring CLV. The tell: you cut channels whose customers repurchase and keep ones whose customers never return.
• Assuming the whole audience is online or on one device. The tell: you cannot say what share of your target is reachable digitally, and the site is not tested on the devices your customers use.

• ▸Acquisition event: state whether it is a first purchase or a contact-permission behavior (and which permission level).
• ▸Target segment: describe behavioral profile (search and email habits, devices), not just demographics.
• ▸Motive evidence: record what research says about why the segment buys, and note if none exists.
• ▸Offline share and other media: estimate the portion of the target not reachable online and the channel assigned to it.
• ▸Channel: name the channel (search, display, vertical network, social ad, email, affiliate, portal, event-driven, viral).
• ▸Destination: record the landing page or conversion path and the persona and purchasing scenario behind it.
• ▸Test design: describe the variable tested (offer, creative, targeting) and the tracking method.
• ▸Cost per acquisition: cost per first purchase or opt-in for the channel.
• ▸Estimated CLV of acquired customers: record the CLV and the cost-to-CLV comparison, then mark keep, change or cut.

1. 1Define 'conversion' in writing for your organization before measuring anything. The source says conversion is organization-specific, so pick the target event: a purchase, a qualified sales lead, or a repeat purchase by an existing buyer. Name one owner for each definition.
2. 2Map the stages between first contact and the target event. For consumers, behavioral data runs 'from information search all the way through to purchase', and search is where most consumers start. For B2B, use the buy cycle and the lead path: generate, qualify, score and distribute, convert. Write each stage as an observable event, such as a search click, a landing-page visit, a form fill, a qualified lead, or an order.
3. 3Build personas and purchasing scenarios for each target segment. Demographics are 'necessary but not sufficient', so attach the behavioral path each persona is likely to follow, such as search then landing page then email, and the device they prefer. Each scenario ends at one conversion event.
4. 4Design a landing page and conversion path for each scenario. The page continues the promise of the ad, search term, or email that sent the visitor. It asks for one next step, and the path from arrival to that step is as short as you can make it.
5. 5Instrument every stage. Use a web analytics tool such as Google Analytics, segment the data (source, campaign, audience), and report conversion rate for each stage transition, not only the final one. A single end-to-end number hides where prospects drop out.
6. 6Test one variable at a time using the direct-response test discipline (Chapter 4). Change the offer, page, or message for a random split of traffic, hold the rest constant, and keep the winner as the new control.
7. 7Treat the first sale as the front end and repeat purchase as the back end. Move first-time buyers into CRM: personalized email and site content, then loyalty programs. Judge whether a conversion is worth its acquisition cost by customer lifetime value, not by first-order profit.
8. 8Review the metrics on a fixed schedule. Check each metric against the goal it was chosen for, and drop metrics that do not change a decision.

The source supplies the framework (acquisition, lead conversion, retention, growing customer value; Chapter 10's lead path; Chapter 14's segmentation and conversion metrics) but no conversion figures, so the numbers below are hypothetical. A B2B software seller defines conversion as 'sales-qualified lead accepted by sales'. The stages are: search or email visit, landing-page form fill, marketing-qualified lead, scored and distributed lead, accepted by sales. Reporting shows 1,000 visits, 50 form fills, 20 qualified leads, and 8 accepted. That is 5% visit-to-form, 40% form-to-qualified, and 40% qualified-to-accepted. The weakest stage is the visit-to-form step, so the team writes a second landing page that matches the search phrase, shortens the form, and splits traffic between the two pages. Qualification and acceptance rates stay unchanged, which tells the team that the new page changed form completions rather than lead quality. After the sale, buyers enter a CRM sequence aimed at repeat purchase, and the team compares channels by lifetime value rather than first-order revenue.

• If you cannot name the single event that counts as a conversion, stop optimizing and define it first. Different teams will otherwise report different 'conversion rates'.
• If the overall rate is low but the stage rates are mixed, fix the largest drop between two adjacent stages first. Do not redesign the whole funnel.
• If a segment converts well on first purchase but rarely buys again, move spend toward retention (CRM, personalization, loyalty) and judge channels by lifetime value.
• If a large share of your target audience is not online, as is true for older, poorer, less educated, and rural segments, set conversion targets for the online channel separately and plan other media for the rest.
• If a test result rests on several simultaneous changes, discard the conclusion and rerun it changing one variable.

• Measuring only the final purchase. The tell: the conversion rate is low but nobody can say at which stage prospects leave.
• Segmenting by demographics alone. The tell: personas read like census rows, with no behavior, device, or search-to-purchase path attached.
• Sending every channel to one generic page. The tell: landing-page copy does not match the ad, search term, or email that drove the click.
• Counting engagement as conversion. The tell: reports cite email opens or time on site while qualified leads and orders stay flat. Engaging in an activity is a different metric from time spent on it, and neither equals a purchase.
• Stopping at the first sale. The tell: acquisition is celebrated, but there is no repeat-purchase rate, no lifetime value, and no post-purchase CRM contact.

• ▸Conversion definition: write the one event that counts as a conversion (purchase, qualified lead, or repeat purchase) and who owns it.
• ▸Segment/persona: name the audience and its behavior (search, email, device), beyond demographics.
• ▸Stage 1 event and count: first observable action (e.g., search click or landing-page visit), with the number for the period.
• ▸Stage 2 event and count: the next commitment (e.g., form fill or add to cart), with the number.
• ▸Stage 3+ events and counts: qualification, scoring, distribution, and purchase, each with its own count.
• ▸Stage-to-stage rates: divide each stage by the previous one and mark the largest drop.
• ▸Test planned: the single variable to change, the control, and the split of traffic.
• ▸Retention step: the CRM action that moves a first-time buyer toward a repeat purchase.
• ▸Lifetime value vs. acquisition cost: record both and note whether the channel pays back.

1. 1Set retention as its own objective. The source's strategy paradigm has four parts: customer acquisition, lead conversion, customer retention, and growing customer value. Write a retention goal separate from acquisition, for example active and repeat-purchasing customers over an extended period, so acquisition results don't mask defection.
2. 2Take the relationship perspective, not the transactional one. The source contrasts the two under Strategic CRM. Decide what you owe a customer after the sale (service, relevant communication, recognition) and who owns each obligation, rather than treating each purchase as a closed event.
3. 3Build the customer database first. The source calls the database an imperative. Record every purchase and every behavioral signal (the clickstream) per customer. Retention programs run from this record, and without it you can't tell who is active, lapsing, or gone.
4. 4Compute customer lifetime value (CLV) by segment. The source places CLV at the opening of its retention chapter and in the direct-response foundations, separating front end (acquisition) from back end (repeat sales). Use CLV to decide how much retention spend each segment justifies.
5. 5Map the customer lifecycle and assign a treatment to each stage. Define the stages (new, active, lapsing, lapsed), the behavioral trigger that moves a customer between them, and one communication per stage. Run the operational side (executing contact, service, offers) and the analytical side (mining the data for who is leaving and why) as two linked processes.
6. 6Personalize with the data you have. Use targeting and personalization in email and on-site content. Use behavioral data from the clickstream for what customers do. Ask them, through marketing research, why they do it, because the source says motivations can only be found by asking.
7. 7Reach customers on the channel and device they prefer. The source says customers have many choices and marketers must satisfy their preferences. Email is used daily by over 90 percent of Internet users and takes more time than search, which makes it a core retention channel. Consider apps and social CRM as additions.
8. 8Add a loyalty program only after steps 3-6 work. Then check cost and failure risk. The source flags CRM system projects as costly and failure-prone, so scope small, tie the system to the lifecycle map, and measure retention and CLV, not just system rollout. The target is a seamless customer experience across touchpoints.

The retention chapter's body is not in the supplied passages. They name Harrah's data mining, the NBA, the International Speedway Corporation, and the ASPCA and MSPCA as cases, but give no results, so none are cited here. What follows is an illustrative walkthrough built only on the source's principles, and its numbers are hypothetical, not from the source. A merchant sees that repeat purchases have stalled. From the customer database, it splits customers into lifecycle stages by days since last purchase and calculates CLV per segment. The clickstream shows that a high-CLV segment has stopped opening email but still searches the site. The data cannot say why, because the source states that motivations come only from marketing research. So the merchant surveys that segment and learns what changed. It then sends personalized content on the channel the segment still uses, and tracks whether the segment returns to repeat purchasing. Success is measured as retained customers and CLV, not email opens.

• If you can't list active, lapsing, and lapsed customers from your database, then build the database before launching any loyalty or CRM program.
• If a segment's CLV is low, then cap its retention spend. Put the budget into the segments where the back end (repeat sales) is worth most.
• If behavioral data shows a drop in activity but not the reason, then run marketing research. Don't guess motives from the clickstream.
• If a customer is active on a channel or device other than the one you use, then move your contact there. Customers' preferences win.
• If a CRM system project is large and expensive, then break it into stages tied to lifecycle treatments, because the source flags high costs and failure rates.

• Counting retention as sales volume: total revenue rises through new acquisition while repeat rates fall. The tell is a healthy top line with no cohort or repeat-purchase data.
• Staying in the transactional mindset: contact stops after checkout, or only promotes the next sale. The tell is a customer record with no post-purchase touches.
• Buying CRM technology before defining the strategy: a system goes in with no lifecycle map. The tell is a live system and no named retention metric, which fits the source's warning on cost and failure rates.
• Inferring why customers leave from clickstream alone: behavior shows what happened, not motive. The tell is offers aimed at an assumed reason that nobody verified.
• Spending equally on all customers: no CLV segmentation. The tell is the same loyalty offer going to your best and weakest segments.

• ▸Segment name: the customer group this plan covers.
• ▸Segment CLV: lifetime value for the segment, with the method and period used.
• ▸Lifecycle stages: list each stage (e.g., new, active, lapsing, lapsed) and its behavioral trigger from the database.
• ▸Preferred channel/device: where this segment engages (email, search, mobile, social), based on observed data.
• ▸Treatment per stage: one personalized communication or service action for each stage.
• ▸Motivation question: what you can't learn from the clickstream, and the research method you'll use to ask.
• ▸Loyalty program (yes/no): include only if the database and treatments exist; note cost and scope.
• ▸Retention metric and target: the repeat-purchase or active-customer measure and the number you're aiming for.
• ▸Review date and owner: who reviews results, and when.

1. 1Define the unit of analysis: one customer's net revenue stream, not one sale. The source's premise is that marketing's job is to create a customer, not just a sale. Mass-media marketers could not do this because they could not identify their end customers. Your first requirement is a way to identify each customer and tie transactions to that person.
2. 2Build the calculation chain in the order the source gives it: net customer revenues, less cost of goods sold, equals gross margin. Subtract the cost of servicing the customer. Then discount the remaining stream by the cost of capital to get the net present value of the customer revenue stream. Keep each line as a separate column so you can see which one drives the result.
3. 3Set the time horizon at three to five years. The source says roughly three years of data are needed to calculate CLV with reasonable accuracy, and after five years the cost-of-capital discount is so high that future revenue has little value. Start collecting customer-level data now, because you cannot build CLV-based programs until the history exists.
4. 4Model the drivers the way the Venkatesan and Kumar figure does. Customer characteristics (switching costs, involvement with the product category or brand, previous purchase behavior) predict future purchase frequency and the profit from each purchase. Supplier-specific factors (channel communication, total marketing communication) set marketing costs. Predicted frequency and profit, less marketing costs, give CLV.
5. 5Use CLV to set the ceiling on spend. The source's logic is that if you know what it costs to acquire, maintain, and service a customer, you can make a reasoned decision about how much to spend to market to that customer. Acquisition budgets should be capped by the CLV of the customer type you are acquiring, not by one transaction's margin.
6. 6Grow CLV through the three levers in the definition. Acquire higher-value customers by targeting on the characteristics that predict frequency and profit. Raise share of customer by increasing purchase frequency and profit per purchase. Retain by treating customer service and experience as part of the offer, as the Ritz-Carlton does by making service culture central and soliciting customer reviews of personal experiences.
7. 7Store the transaction, service-cost, and contact-cost history in a customer database. Use it for program execution and analytics, with profiles and predictive models built on top, so CLV is recalculated as behavior changes rather than set once.
8. 8Test before rolling out. The chapter pairs CLV with testing (A/B split tests) and an offer. Test acquisition and retention offers against their effect on CLV, not only on the first response.

The passages do not reproduce the chapter's own numeric CLV example, so this walk-through follows the source's structure without inventing figures. Take a luxury-hospitality customer like the Ritz-Carlton's, where customer service is part of the overall customer experience and the firm asks guests for reviews of their stays. Step 1: sum the customer's revenues across expected stays. Step 2: subtract cost of goods sold to get gross margin. Step 3: subtract the cost of servicing, which for a service-intensive brand is large and must be counted. Step 4: subtract marketing communication cost per the Venkatesan and Kumar figure. Step 5: discount each year's remainder by the cost of capital across a three-to-five-year horizon and sum to get net present value. The result is the most you can rationally spend to acquire and keep that customer. Predicting frequency from switching costs, involvement, and past behavior tells you which guests deserve the higher spend. Higher service cost is justified only if it raises retention or frequency enough to lift the discounted total.

• If you have less than about three years of customer-level data, then treat any CLV figure as provisional and avoid committing large acquisition budgets to it.
• If a cost to acquire a customer type exceeds that type's CLV, then cut or restructure the acquisition spend, however good the first-sale margin looks.
• If the projected stream extends beyond five years, then ignore the tail, because the cost-of-capital discount leaves it with little value. Plan on three to five years.
• If customers show high switching costs or strong involvement with the brand or category, then forecast higher purchase frequency and invest more in retention contact.
• If you cannot identify individual customers, then fix identification and the customer database first, because without it you cannot interact directly or calculate CLV.

• Treating CLV as revenue, not net present value: the tell is a CLV figure with no cost of goods, servicing cost, or discounting lines.
• Judging campaigns by the first sale: the tell is acquisition reports that stop at conversion and never track later purchases.
• Calculating on too little history: the tell is a precise CLV built from a few months of data, when about three years are needed for reasonable accuracy.
• Counting an overlong horizon: the tell is projected revenue beyond five years inflating the total, even though the discount makes it nearly worthless.
• Leaving service cost and experience out of the model: the tell is a retention strategy with no servicing-cost line and no customer feedback loop, such as reviews.

• ▸Segment definition: describe the customer type and how you identify its members.
• ▸Horizon (years): enter 3 to 5 and state how many years of actual data you hold.
• ▸Net customer revenues per year: list expected revenue for each year of the horizon.
• ▸Cost of goods sold: subtract per year to reach gross margin.
• ▸Cost of servicing: enter the annual cost to service and maintain this customer.
• ▸Marketing communication cost: enter the annual cost of channel and total marketing contact.
• ▸Cost of capital (discount rate): enter the rate used to discount each year's net amount.
• ▸Predicted drivers: record switching costs, involvement, and past behavior, and the purchase frequency and profit they imply.
• ▸Net present value (CLV): sum the discounted yearly nets; this is the spend ceiling for acquiring and keeping the segment.

1. 1Start from objectives, not channels. The source's rule is that every channel, electronic and physical, 'must be driven by clear marketing objectives and monitored for effectiveness.' Write one objective each for acquisition, conversion, and retention before you assign any channel.
2. 2Inventory every channel you run, split into offline (physical world: print, radio, events, kiosks, stores) and interactive (website, blog, search/SEO/PPC, email, mobile, social). Treat the Internet as one channel among many, not a stand-alone medium.
3. 3Assign each channel a job by funnel stage. The text places search as the single most important way to drive visitors to a website, and email as usable for acquisition but especially important for retention. Offline channels get a stage too, and a hand-off point to an online destination.
4. 4Write one core message and one set of brand cues, then check each touchpoint against it. A customer who sees the offline ad, the search result, the site, and the email should be able to tell it is the same brand making the same promise.
5. 5Connect the channels through a shared customer database. Database foundations are the basis of Internet direct response, and the CRM vision in the source is a 'seamless customer experience.' Every channel should read from and write to the same customer record so one channel knows what another already did.
6. 6Design the hand-offs explicitly. For each offline touchpoint, specify the online action it points to (URL, search term, email sign-up). For each online touchpoint, specify the offline or service option it can route to. Test each hand-off as a customer would.
7. 7Compare cost and results by channel, including customer service. The source treats cost of customer service by channel as its own topic, so a channel that is cheap to run can still be costly to serve through.
8. 8Reallocate budget from measured results, not trend. Review monitoring data on a fixed cycle and shift spend between online and offline channels according to which ones advance the stated objectives.

The source's B2B budget data (eMarketer/ExactTarget, 2010) shows how channel mix is shifting. 70% of responding companies planned to increase budgets for off-site social media. Only 17% were increasing print budgets, against 41% decreasing. 15% were increasing radio, while 36% were spending less. More than half planned to increase mobile (56%), email (54%), and paid search (51%). The passages do not give a single company case of integration, so this shows how to read the data. A B2B marketer seeing these shifts should not simply move money out of print and radio. Following the text's rule that online and offline must be integrated and monitored against objectives, she first lists what print and radio do for acquisition. She then checks whether the growing channels (search, email, mobile, social) can carry the same message and capture the same leads through the shared database. Money moves only where the monitoring data show the objective is still met. The data are a market signal about peers, not evidence about her own customers.

• If a channel cannot be tied to a stated acquisition, conversion, or retention objective, then do not fund it until it can.
• If an offline touchpoint creates interest, then give it a trackable online destination, because search is where most consumers start the purchasing process on the web.
• If a message or offer differs between channels, then fix it before adding spend. The goal is a consistent, seamless experience.
• If peers are cutting a channel (for example print or radio), then test your own results before following them. Reallocate only where your monitoring supports it.
• If channels do not share customer data, then fix the database before adding channels, because the CRM vision depends on one view of the customer.

• Treating the Internet as a stand-alone channel. The tell: a separate online team, budget, and message with no link to offline plans.
• Following industry budget trends instead of objectives. The tell: you can cite what peers are cutting but not what a channel contributes to your acquisition, conversion, or retention goals.
• Channel silos with separate customer records. The tell: a customer gets an acquisition offer for something already bought, or has to repeat information when moving from one channel to another.
• Inconsistent brand message across touchpoints. The tell: the ad, the landing page, and the email promise different things.
• Never monitoring effectiveness by channel, including service cost. The tell: budget moves by habit and you cannot say which hand-offs lose customers.

• ▸Objective: state the acquisition, conversion, or retention goal this map serves.
• ▸Channel: name each offline and interactive channel (e.g., print, radio, events, kiosks, website, search, email, mobile, social).
• ▸Online or offline: label the channel type so you can see the balance.
• ▸Funnel role: acquisition, conversion, retention, or service.
• ▸Core message check: does this touchpoint carry the same brand promise? (yes/no, plus fix).
• ▸Hand-off: where does this channel send the customer next, and by what mechanism (URL, search term, sign-up)?
• ▸Data captured: what customer information goes into the shared database from this channel.
• ▸Cost and result metric: spend and the measure tied to the objective, including service cost if relevant.
• ▸Budget decision: increase, hold, or decrease, with the evidence for it.

1. 1Treat every mouse click as a candidate data point. List each interaction on your site, email and ads (page view, click, registration, order, service request) and mark which ones you currently capture. Anything uncaptured is lost, because only marketers who capture and use this data can tap the full power of the interactive web.
2. 2Define 'the marketing database' correctly. It is not one file. It is the customer database plus several other databases, all linked through a central repository such as a data warehouse, so information reaches decision makers and operations staff on demand. Draw your own version of the Figure 4.9 map: each database, the data it holds, and its primary source.
3. 3Sort your databases into the four warehouse components: (1) marketing management databases that drive programs, (2) marketing support databases that inform decisions, (3) functional-area databases such as order processing, inventory, sales force management and projects, and (4) externally purchased data. Mark which of these are not yet linked to the warehouse.
4. 4Link order-processing and inventory data to the customer record first. The source calls both essential to good customer service, so marketing promises should reflect what can actually be fulfilled.
5. 5Use the database for program execution, not just reporting. Feed it into the live customer experience: recommendations based on preferences and prior purchases, order history, customer-service request history, and customer-set communication frequency preferences.
6. 6Use the database for analytics, including data mining. Mine the warehouse for patterns that change targeting and program decisions. The book's data mining examples are Harrah's and the NBA (the passages give only the headings, so read those sections for detail). Tie each analysis to a decision it will change.
7. 7Tie data use to customer lifetime value and testing. The book places CLV and direct response testing alongside the database imperative. Segment, personalize and test by the value you expect from each customer, not by cost per click.
8. 8Treat the database as institutional memory. The source notes it can partly offset the churn of human resources, so document field definitions, sources and the logic behind each segment so the knowledge stays when staff leave.

The passages describe an unnamed company whose customer service and marketing run on one customer data foundation. A registered customer sees recommendations based on preferences and prior purchases. She sets her own communication frequency preferences and can view her order history and the history of her customer service requests and responses. The company also runs an online loyalty program with free shipping and 5 percent cash back, plus an active web community. Because it supports multiple channels (customers often use several to get results) and integrates experiences within the website, the same data serves service and marketing. The reported results are a Net Promoter score up nearly 7 times, customer satisfaction scores up 10 percent, and email volume cut by 72 percent, with customer service managed more cost effectively. The lesson for your own build is to capture preference, purchase and service-history data in one place, expose it back to the customer, and measure the effect on service cost and satisfaction. The passages do not say which databases or tools the company used, so treat the data design as your own work.

• If a data source (orders, inventory, service history, sales force, purchased data) is not linked to the central repository, then connect it before building new personalization, because decision makers cannot get information on demand from siloed files.
• If a click, registration or transaction event is not captured, then fix capture before analysis, since every mouse click is a possible data point and uncaptured clicks cannot be mined.
• If an analysis will not change a targeting or program decision, then drop it, because the database exists to support better marketing decisions.
• If customers use multiple channels to get results, then expose one history (orders and service requests) across channels rather than one per channel.
• If you offer personalization, then let customers set their own communication frequency preferences and honor them.

• Treating 'the marketing database' as one customer file. The tell: order, inventory and sales data live elsewhere and no one can answer a customer question from a single view.
• Collecting clickstream data without a repository or use case. The tell: large log volumes, but no segments, tests or decisions trace back to them.
• Using the database only for reporting, not for execution. The tell: the website shows registered customers generic content despite known preferences and prior purchases.
• Ignoring service data in marketing. The tell: customers get promotions while an unresolved service request sits in another system, and email volume stays high.
• Depending on staff knowledge instead of the warehouse. The tell: when an analyst leaves, no one can explain how a segment or report was built.

• ▸Data source: name each database (customer, order processing, inventory, sales force, project, support, purchased).
• ▸Warehouse component: assign it to marketing management, marketing support, functional area, or external.
• ▸Primary source: state where the data originates (web click, registration, order, service, vendor).
• ▸Linked to central repository? Yes or no, with the owner and a date to connect it if no.
• ▸Click and transaction events captured: list the web interactions currently recorded as data points.
• ▸Events not captured: list the interactions you are losing and the fix.
• ▸Execution use: name the customer-facing use (recommendations, order history, service history, frequency preferences).
• ▸Analytical use: name the question being mined (data mining or CLV) and the decision it will change.
• ▸Result metric: state the measure to track (Net Promoter score, satisfaction, email volume, service cost) and its baseline.

1. 1Start from the strategy objective, not the tool. The source frames every tool against four goals: customer acquisition, lead conversion, customer retention, and growing customer value. Write one objective per campaign in those terms before you pick a channel.
2. 2Name the target audience and where it spends its time online. The passages say consumers spend a large share of their Internet time on social networks (communicating, getting peer information, occasionally buying), and B2B customers use social media for collaboration and information. Match the channel to the audience's behavior, not to what is familiar.
3. 3Assign each tool a job. Search is where most consumers start the purchasing process, so cover both natural search optimization and pay-per-click. Email remains a key tool for B2C, B2B, and nonprofits. Display advertising serves branding or direct response, and you declare which one per placement. Social media handles peer-based information and collaboration. Video and online branding build the brand.
4. 4Check the format before you buy display. The source says display is undergoing a renaissance as new formats engage the viewer better. Choose the format that fits the objective (brand message or response action) instead of defaulting to a standard unit.
5. 5Account for mobile. The source notes consumers continue to migrate to mobile communications and text messaging, even as email stays essential. Confirm that each email and landing experience works for someone arriving from a phone.
6. 6Integrate across media. The source says the Internet is no longer a stand-alone medium and that strategies and messages should be integrated across online and offline channels. Write one core message and specify how each tool carries it, including any offline touchpoints.
7. 7Pre-test the destination. Usability testing is a pre-testing technique that checks the site works in a user-friendly way for members of the target market. Run it before launch or redesign, and again if other metrics suggest site problems. Keep it separate from testing of communication appeals, which belongs to the marketing communications program.
8. 8Define measurement per perspective before launch. Figure 14.2 separates traffic, audience, and campaign measurement from site performance (quality assurance) and usability. Assign each tool a metric in the right category so you do not use traffic numbers to judge usability.

Amazon.com, as described in the passages, shows tools deployed against retention and customer-value objectives on its own site. The product lines (books and similar standardized goods) are not distinctive, so Amazon uses tools to make the experience nonstandard and to encourage repeat transactions. It built a search engine based on Google functionality that displays a richer set of information, because a huge book selection made search essential. It used collaborative filtering (software that finds patterns of activity) to personalize recommendations. Those recommendations are visible on a personalized home page and in the 'My Store' feature, and the customer can edit purchase history or rate owned products to sharpen them. It also used customers to supply some site content, and it patented 1-Click, which remembers purchase and mailing information so the customer skips lengthy forms. The pattern: each tool serves a defined objective (findability, relevance, low friction), and each is tied to customer data. The passages give no campaign figures for this, so treat it as a model of deployment logic, not a benchmark.

• If the objective is acquisition and the audience starts its purchase journey by looking things up, then put search first, covering natural optimization and pay-per-click.
• If you have identified existing customers in a database, then use email and event-triggered communication (a purchase or a birthday starts an automated message) for conversion and retention.
• If the display placement is for branding, then judge it on brand measures. If it is for direct response, then judge it on response. Never mix the two in one report.
• If the audience is B2B, then include social media for collaboration and information gathering, and do not treat it as consumer-only.
• If other metrics suggest site problems, or you are redesigning, then run usability testing before sending more traffic to the site.

• Tool-first planning: you buy a channel because it is fashionable. The tell is that no objective from the acquisition/conversion/retention/value set is written next to it.
• Treating the Internet as a stand-alone medium. The tell is that online messages differ from offline ones, or offline channels are missing from the plan.
• Dropping email because consumers are moving to mobile and text. The source says email stays key, so the tell is an email program shut down without a mobile-ready replacement.
• Sending traffic to an untested site. The tell is rising traffic or campaign clicks with poor conversion, and no usability test on record.
• Mixing display branding goals with direct-response goals. The tell is one placement reported against both metrics, so neither result can be read.

• ▸Objective: choose one of acquisition, lead conversion, retention, or growing customer value, and state it in a sentence.
• ▸Target audience: describe who they are (consumer or B2B) and where they spend online time.
• ▸Tool selected: search (natural/PPC), email, display, video, social, or mobile/text.
• ▸Tool's job: state what this tool does that the others do not.
• ▸Display format and purpose (if used): name the format and mark branding or direct response.
• ▸Core message and offline tie-in: write the one message and list the offline touchpoints that repeat it.
• ▸Destination and usability test: name the landing page and record the test date and result.
• ▸Metric and perspective: pick one metric each from traffic/audience/campaign and from site performance or usability.
• ▸Review trigger: state the result that causes you to change or drop the tool.

1. 1Set the permission level before you design any content. The source contents list 'Levels of Permission Marketing' as a distinct topic in the email chapter, so decide per program how much consent you hold (for example, a bare address versus a stated-preference profile) and write that level next to the program. Personalization depth cannot exceed the permission behind it.
2. 2Make the opt-in explicit and tied to a stated benefit. At signup, say what the person will receive and ask for preferences (topics, product interests, frequency) rather than only an address. Stated preferences are the cleanest personalization input because the customer supplied them.
3. 3Build the profile from two layers: what the customer told you (preferences, needs) and what their activity shows. The passages note that user activity on the web creates databases that are 'the basic resource of Internet marketing strategies.' Keep the two layers labeled separately so you know which personalization rests on consent and which on inference.
4. 4Turn profiles into segments with customer personas and purchasing scenarios. The lead-generation material lists 'Creating Customer Personas' and 'Developing Purchasing Scenarios' as separate tasks. Write each persona's need, the scenario in which they buy, and the one message that fits that scenario.
5. 5Climb the hierarchy in order: targeting, then personalization, then customization. The contents list 'The Hierarchy of Interactive Strategies' and 'Targeting, Personalization, and Customization' as separate ideas. Start with targeting (right segment), add personalization (content tailored to the identified person), and offer customization (the recipient adjusts what they get) wherever recipient control matters.
6. 6Apply personalization to both email and site content. The CRM material lists 'Personalized Email and Site Content' together. Keep the offer, the language, and the on-site experience consistent for the same identified customer.
7. 7Run a compliance and privacy check before every send. The email chapter covers the U.S. CAN-SPAM Act and its application in the European Union. The privacy chapter covers consumer attitudes toward privacy and the Fair Information Practices Principles. Confirm consent status, include a working way to opt out, and honor it.
8. 8Track results and feed them back. The email chapter includes 'Tracking Emails.' Compare personalized with non-personalized sends, check opt-out and complaint signals, and revise segments and preference options before you widen the program.

The source passages name Scott's Turf Builder and Dreamfields Pasta as email marketing cases, but they supply only the titles, so no figures or tactics from those cases are used here. This is an illustrative build, not a reported result. A lawn-product marketer collects opt-ins on its site with a one-line promise ('seasonal lawn-care tips') and a preference field for lawn type. Customers who give a lawn type form the segment 'new-lawn owners.' That segment receives a sequence tied to its purchasing scenario (seeding, then feeding) instead of the general newsletter. The marketer adds a preference link so recipients can change topics or frequency, which is customization under the recipient's control. Each send includes the required opt-out and sender identification, and the marketer tracks opens, clicks, and unsubscribes by segment. If unsubscribes in a segment exceed the general list's, the marketer cuts the frequency before adding more personalization.

• If you hold only a bare address with minimal consent, then limit personalization to broad segment targeting and ask for preferences in the first message. Do not infer sensitive needs from behavior.
• If a recipient states a preference that conflicts with what behavioral data suggests, then follow the stated preference. They told you; the inference is a guess.
• If a segment's personalization is built from inferred data, then keep the message general enough that a wrong inference reads as harmless, and offer a link to correct the profile.
• If opt-outs or complaints rise after you add personalization, then reduce frequency or depth and give more recipient control (preference options) before adding more tailoring.
• If a program crosses borders, then check the privacy and anti-spam rules of each country, because the source treats CAN-SPAM in the EU and privacy regulation by country as separate matters.

• Treating a purchased or inherited list as opted-in. Tell: low engagement and complaints from people who do not recognize the sender.
• Personalizing with data the customer did not knowingly provide. Tell: replies or complaints asking how you knew that, and a rise in opt-outs after the personalized send.
• Using the first name as the only personalization. Tell: segment-level results match the generic send, because the content ignores the persona's need or purchasing scenario.
• Offering no way to change preferences, only to unsubscribe. Tell: recipients leave entirely instead of reducing frequency or changing topics.
• Skipping the compliance check on cross-border sends. Tell: the program is built to U.S. rules and nobody has reviewed the other jurisdictions' requirements.

• ▸Program name and audience: one line naming the list and who it serves.
• ▸Permission level held: state exactly what the customer agreed to and how it was captured (form, date, wording).
• ▸Promise made at opt-in: write the benefit stated to the customer; every send must match it.
• ▸Stated preferences collected: list the fields customers fill in (topics, interests, frequency).
• ▸Behavioral data used: list each inferred signal and mark it 'inferred' so it stays separate from stated data.
• ▸Persona and purchasing scenario: name each segment, its need, and the situation in which it buys.
• ▸Personalization depth: mark each element as targeting, personalization, or customization.
• ▸Recipient controls: describe how people change preferences, reduce frequency, and opt out, and where the links appear.
• ▸Compliance and privacy check: record CAN-SPAM and any non-U.S. rules reviewed, the reviewer, and the date.
• ▸Result review: list the metrics tracked per segment (opens, clicks, opt-outs, complaints) and the change you will make if opt-outs rise.

1. 1Define engagement as observable behaviors before you pick channels: attending (time spent on a post), interacting (responding to campaigns or offers), sharing, and collaborating with the brand. The passages tie engagement to these states, so each one needs its own measure.
2. 2Start from behavioral data, not demographics alone. The source says demographic data are 'necessary but not sufficient' and that behavioral data are key from information search through purchase. Segment your audience by what they do (email, search, social, video), then check demographic gaps. Older, poorer, less educated, and rural segments are less likely to be online, so some of your audience may need other media.
3. 3Match the channel to the audience. In the passages' marketer survey, Facebook is the most effective channel for reaching consumers. LinkedIn, blogs, and Twitter are the effective channels for reaching business people. Put engagement effort where your target segment already spends attention.
4. 4Embed richer, interactive content in the places people already read. Embedding videos or slide presentations in blog posts increases engagement. The source author tells which posts carry embedded media by the time spent on the post.
5. 5Use interactivity where the evidence supports it. Interactive ads in a digital magazine produced greater perceived interactivity, perceived engagement, message involvement, and a more positive attitude toward the ad than the print edition. Where you can, build content that people act on (unfurl, click, register) instead of only view.
6. 6Fit the engagement tactic to the stage of the consumer decision journey. Company-driven marketing such as advertising has the most impact at initial consideration. Past experience dominates evaluation. Word of mouth and in-store interactions dominate closure. Use ads to attract attention, then build shareable and collaborative touchpoints for the later stages.
7. 7Connect content across channels. Tweets link to videos and to blog posts that embed videos, so the content ecosystem is interconnected. Plan each asset to point to the next one, and reach customers on whatever device they prefer.
8. 8Build a loop for the brand relationship. A community model rests on user loyalty and repeat engagement, with users investing time and emotion. Give people a way to contribute, share, and return, and track repeat behavior, not only one-time reach.

Martha Stewart Living launched its iPad app in November 2010 with an entire issue created for the device. On her blog she posted a cover shot and a demonstration video, and the cover could be watched 'unfurling' a peony built from 180 time-lapse photos taken over 10 hours in her garden. The interactivity drew praise from app critics and CNBC. The support page also asks readers without an iPad which devices they use, which invites customers to say how they want to engage. The supporting study found interactive digital ads produced greater perceived engagement and message involvement than print. By February 2011 the magazine itself was published digitally. The pattern to copy: interactive content, a blog demo to pull attention, and a direct request for customer input.

• If your target segment is business buyers, then prioritize LinkedIn, blogs, and Twitter. If it is consumers, then lead with Facebook.
• If a post earns little time on page, then embed a video or slide presentation and compare time spent before and after.
• If the goal is initial consideration, then use company-driven marketing like advertising. If the goal is closure, then invest in word of mouth and in-store or direct interactions.
• If demographic data show part of your audience is offline (older, poorer, less educated, rural), then reach that portion through other media instead of expecting digital engagement.
• If a premium or interactive format is available, then keep it uncluttered with run-of-the-mill advertising, since engaging formats are not optimally effective without strategy at the right stage.

• Treating reach or demographics as engagement. The tell: reports list audience size and age bands but no behavior such as time on page, responses, or shares.
• Spreading content across every channel regardless of audience. The tell: effort on Facebook while your buyers are business people who respond on LinkedIn, blogs, and Twitter.
• Using engaging formats at the wrong journey stage. The tell: interactive ads are pushed for closure, when word of mouth and in-store interactions matter most there.
• Publishing isolated assets. The tell: videos, tweets, and blog posts do not link to one another, so the content ecosystem is fragmented.
• Assuming one device or format. The tell: content is built for one platform, such as an iPad-only issue, with no question to customers about the devices they use.

• ▸Target segment: describe the group by behavior first, then demographics.
• ▸Offline gap: note which part of the segment is not online and the other media you will use to reach it.
• ▸Primary channel: pick the channel where this segment is most reachable (e.g., Facebook for consumers, LinkedIn/blogs/Twitter for business).
• ▸Journey stage focus: initial consideration, evaluation, or closure.
• ▸Content asset and interactive element: name the asset (video, slides, interactive ad) and what the customer can do with it.
• ▸Linked assets: list the posts, tweets, or videos this asset points to or embeds.
• ▸Engagement measures: define time spent, responses, shares, and collaboration actions, with a baseline for each.
• ▸Customer input mechanism: describe how customers tell you the device or format they prefer.
• ▸Review result: compare measures to baseline and note the next change.