Online to Offline Conversions – A modern example from National Instruments

(This post is part of a series on the state of multichannel metrics today, one year after the book came out.)

The previous post paid homage to the original innovators behind the online-to-offline conversion improvement idea, namely the automobile industry. Now, let’s look at a modern adaptation from the B2B, high tech, sector at National Instruments.

National Instrument’s Michelle Rutan and I presented this case together at the eMetrics Marketing Optimization Summit in San Jose this May. By popular demand, the session will be repeated live, online, namely in the WAA”s upcoming webcast on July 21st.

The challenge

Take these stats in:

  • The company’s web site has 1 million unique URLs.
  • Visitors reach the site through one of 500,000 monthly keywords.

Add to that the fact that there are more than 6,000 part numbers of National Instruments (NI) products. So what we have at hand is a great challenge for managing a long, long tail of web pages, keywords, and products.

In this complex situation, how can the sales and marketing team get insight into what prospects are interested in? Imagine being a sales person and asking your account whether they’d be interested in one of 6,000 products today?

Hey, how about a 200 MhZ Wave form generator for you today? Would you like Fast, Flexible 6.6 GHz RF Instrumentation with that?

Or imagine being a marketing person crafting an email that will be perceived as relevant.

Hey, we got to Wave form generators for the price of one today?

Boy, that seems hard.

The general idea

Much like car buyers, many of NI’s buyers use the web site to do their own product research. So NI said to themselves, why should we ignore the click signals that buyers are sending us through their keyword searches and content views?

As a side note, there is one great advantage for this in B2B that some companies leverage. Namely, buyers need not even be registered for their clicks to be attributed to a specific prospect. Often it suffices to translate the IP address into the company’s domain name to know that somebody from company X has been visiting the site.

“What happens on the web site, stays on the web site”, no more.

You could say that in web marketing 1.0 the web site was a silo by itself. Prospects might do their research but marketers and sales teams would pretty much ignore them until they email or call in.

At NI, the marketing team has been pondering for a long time already how they might turn the data on visitors’ behavior into better customer service.

  • Surely, a recent and frequent visitor to the product (not support) section of the site is more likely to be an active prospect.
  • Surely, a visitor who is deeply engaged with product pages (again, not tech support pages) is likely to be an active prospect.

But identifying active prospects is only part pf the equation. With a long tail of products the question for sales and marketing remains what these prospects are interested in.

Getting to the essence of a prospect’s interests

Profiling visitors’ interests based on their web behavior turned out to be much easier said than done.

NI experimented with several methods of mapping keywords and web content to the products that the pages relate to. Turns out this isn’t so easy at all. Several methods were tried and eliminated.

  • Do each of 1 million URLs really contain the necessary information for mapping them to their content?
  • Can you conceive of a way to map that many URLs to their essential contents manually — and keep the mapping up to date?

Not so easy.

Then Michelle Rutan had another totally cool idea of how content and keywords could be mapped out. Applying this new mapping to her web analytics data, she generated a test segment of prospective buyers for a particular product and geo area.

The Results

The test campaign was launched as an email and achieved a whopping 44% open rate and 25% click-through rate!

This is great testament to the fact that recipients perceived the content as relevant.

The data is still so fresh that resulting sales information was not available (not that it would be publicly available anyway).

The million dollar question:How did Michelle map web behavior to prospects’ essential interests?

Since this is entirely Michelle’s idea, it isn’t for me to reveal her innovative approach. But if you are curious, you can hear from her directly by tuning into the upcoming webcast.

Michelle will present from the business perspective. Then I will run through a lab case of how this idea can be turned into reality by integrating web analytics into the data warehouse, BI, or CRM system.

I know, Michelle and I look forward to your questions.

Replay of May 19th Webcast with Kevin Hillstrom and Jim Novo

If you missed the May 19 WAA Webcast with Kevin Hillstrom and Jim Novo, you can replay it any time on demand.

By the way, do you think it will be 5 or more like 10 years before all TV will be much like the Internet?

That is to say, you will turn on the tube and a big bing or google box will appear in the middle of the screen. You’ll type in “Kevin  & Jim webcast”  and get your multichannel marketing fix while sipping a cup of old fashioned tea.

Unless, of course, you see my PPV (pay-per-view) ad show up towards the right of your TV screen and click on it to read this blog.

Meanwhile,  recommendations will appear at the bottom of the screen that are targeted to your remote control behavior.

Hopefully, something better than “Meet exciting online and offline marketers in <your city>”. 😎

April 23, Santa Clara: Managing Integrated Marketing

If you liked the previous post, then you will love this.

Well, only if like me you live in the SF Bay Area, that is.

Friends of analytics are invited to some quality evening entertainment on Thursday, April 23d, at 5.30 pm. The Silicon Valley AMA is conducting a panel discussion in Santa Clara. The topic: Managing integrated marketing through analytics.

The event is generously sponsored by one of my most highly regarded competitors: Coremetrics.  And Coremetrics have been even more generous by permitting keynote speaker Andreas Ramos, co-founder of analytics agency The CCG Group, to invite me to join the panel.

As for me, I am burning of course to learn the kind of questions that the audience will ask and to learn from the experience of the other panelists.

If you are local, come and join! Register here.

Thank you to Andreas for the invitation.

Competing of Data … for Competing on Analytics

For those about to compete on analytics – I salute you!

The reference to ACDC isn’t unfounded. After all … data rocks!

If you want to create competitive advantage for your company by using analytics more cleverly than your competitors, then the first question is:

What kind of data will you use for those analytics?

When it comes to customer data, most marketers today have many choices to to pick from. Their (unified or disparate) data marts contain everything from

  • transactions
  • to personal details
  • to marketing contact and response history,
  • and in much too few cases also the web site interaction history of individuals, i.e. web analytics at the personal level

Leaders have found ways of using some of this data to leave the competition in the dust:

  • Online marketers may think of Amazon, for example, or its cousin in the Netherlands:
  • Offline marketers may think of companies such as CapitalOne who grew rapidly thanks to clever direct marketing fueled by analytics (as documented in the book Competing on Analytics.)

Yet at many other companies the customer data unfortunately are sitting idle in the data mart, collecting dust, and are not getting leveraged as well as they could be.

With some imagination, you could visualize the idle data silos talking among themselves and scheming how to get out of their isolation and boredom.

You could say, the data are in a competition with each other to get adopted by the marketer first. For that, each data source needs to make the case that it is the best weapon for the marketer to take with her into the battle against the competition.

“Hey, pick me, I can help you more than the next data silo”

So in the hustle and bustle of the various data sources fighting it out with each other, you might catch following battle cries:

“I am the transaction data”

I come in many forms, for example, shopping baskets at retailers, call data records at Telcos, and account transactions at banks or credit card companies. Since virtually all companies study me already though, it is going to take some more ingenious analytics before you can differentiate yourself using me.

Such distinguished analytics can for example be behavioral event detection, i.e. the detection of changes in individuals’ patterns of behavior. For example, banks do this very successfully by flagging individual customers who may be ready for cross-sales or retention efforts. This may for example be the case when a customer has an unusually large deposit on their bank account relative to the individuals’ personal past deposit averages.

Companies that use me well have increased their marketing success rates 5 to 12 times. So you betcha you can compete using me!

“I am the marketing contact and response history”

Most marketers think of me as just a tactical “log” of past interactions. If at all, I am used for calculating a direct marketing campaign’s response rate.


But in today’s world of multichannel, interactive (or dialog) marketing, I have a much more strategic role to play. Namely, a marketer that doesn’t take me into account is like someone who is talking while turning a deaf ear to the conversation partners’ responses.

Therefore, I am the one who can help you go from mass marketing to interactive (or dialog) marketing. Since few marketers use me well today, you can really compete on analytics with me!

“I am demographics data”

Most companies have some form of me available.

I am often helpful for framing who is in the company’s target audience so that you don’t waste your marketing funds talking to people who won’t buy anything from you anyway.

But I have been around for such a long time, even I can’t remember how you could use me to create a competitive advantage. What can you do with me that your competition isn’t already doing?

“I am the customers’ permission and preference data”

I may seem like a bore at first — after all I represent the customer’s interests and not the marketer’s. But companies that use me well are able to continue the interactive exchange with customers while companies who ignore me lose their permission to market.

For example, leaders in email marketing offer their subscribers not just an “opt-out” but a way to manage for themselves when and how often they wish to be contacted about what. Ditto with RSS marketing. Instead of losing a prospect to an opt-out, you are given a chance to be relevant.

“I am the web analytics data”

Most web marketers package me into reports and good looking dashboards. They use me to make their web sites and advertising more successful.

But since I am used that way by almost everyone already, it is really tough for you to compete on me this way.

You have to be cleverer than that to turn me into gold!


A straight forward but rarely tapped opportunity is to make web analytics personal (download the whitepaper), i.e. to learn about each individual prospect or customers’ current interests as demonstrated by their most recent web site sessions, clicks, and keywords.

This can fuel behavioral targeting, event based marketing, or highly predictive analytics. Marketers can use me to send the right re-marketing, on-boarding, cross-sales or retention recommendation to each customer at the right time.

Companies that use me well today are the leaders in their space, e.g. Amazon, eBay, Verizon, and many others.


So, if you have so many data to choose from which should you pick?

This will depend on your business and your competition. Most likely you have many more than just a single opportunity. And just like any other marketing investment you want to forecast potential returns vs. costs of getting there.

You’d start implementing the opportunity that has the best potential. But you shouldn’t stop there. Rather, all initiatives that promise a lucrative ROI (above your hurdle rate) are worth doing and should be funded. That is the only way in which you will maximize total returns.

So, go ahead, make your business case and your CFO will get you the funds.

Yes, these days many budgets have been cut.

But just today I was hearing from a business intelligence manager at a large client of mine. She made her case for solving a long standing business problem through extremely innovative use of (web) analytics. This was a business problem that the company hadn’t been able to solve through any other means. And sure enough, three weeks ago she got the resources that she wanted.

For those about to compete on analytics, we salute you. You rock!

A nugget from our webinar with Bill Leake from Apogee-Search

Here is a quick take-away from our webinar with Apogee’s CEO Bill Leake today: Don’t be shortsighted in applying analytics for boosting success from online lead generation, specifically paid search:

  • Don’t just measure conversion rate. It is a tricky metric that can often lead in the wrong direction. Instead be sure to measure further down the funnel, as Bill  emphasized
  • Measuring further down the funnel means getting closer to ROI metrics, e.g. a cost per sale or cost per lead.
  • Ideally you measure even further down the funnel to get to an ROAS or ROI metric.
  • If you are selling offline rather than online then you can do this by adding the data on individual’s keyword referred visits into the CRM system

Continuing on the theme that Bill started, I also recommended using web analytics in more than just shortsighted ways:

  • Don’t just stop at reporting results
  • Don’t even stop at improving results by making informed changes to fix funnel leaks indicated by your web analytics
  • But open up the gold mine of making web analytics personal, i.e. behavioral targeting or what we refer to as Interactive Marketing at Unica.
  • For example, if Mary searched for blue shoes coming to your web site then:
    • Of course the web site should take that into account in targeting content to Mary.
    • But your email marketing can leverage the same info to personalize its outreach
    • And there is no need to forget about Mary’s interest in blue shoes when you send her an offline brochure or she calls your call center. Rather, today’s marketing automation solutions enable you to extend the dialog beyond  online to the  offline channels.

If you missed the event, here is a replay link for the webinar with Apogee and Unica.

Thanks for all the attendees that dialed in. Bill and I will post the question to which we couldn’t get shortly.


Need to Print Money? Learn how, on Feb 18 with Apogee and Unica

The value prop of being better at SEM is that you can achieve the same outcomes with less spending.

  • Reduce spend on PPC keywords
    • that attract visitors who would have clicked on your organic search result anyway
    • Reduce spend on PPC keywords that don’t carry their weight
  • Lift the results for all PPC and organic keywords
    • by fixing the leaks on the web site
    • through intelligent targeting, lead-nurturing, and re-marketing that takes into account the keywords through which prospects spelled out their interests

Come, join Bill Leake, a former McKinsey & Co consultant and CEO of Apogee Search, and myself on a webinar, Feb 18, where we will fill these bullets with life. Ping us with your own questions and keep us real.

Register here.

Feb 5th – WAA Webinar: Get the $$$ for Testing

If you had the chance to put together an educational webinar on the topic of testing website content and ads – who would be your dream picks as panelists?

Well, I was super lucky to be in just that position to propose a panel for the topic.

And I think we are all super lucky that testing gurus Bryan Eisenberg from Futurenow Inc. and Josh Manion from Stratigent agreed to be the panelists in the upcoming Web Analytics Association webcast:

Testing, Testing, Testing
Anybody Can test.
So Why Don’t You?

Who are these guys?

You will know Josh Manion from his work at Stratigent, his frequent presentations at emetrics, and educational seminars. Stratigent have tremendous experience with testing on behalf of their clients. They employ various methods for testing including for example the multivariate testing solution Optimost from Interwoven. (acquired by Autonomy a few days ago).
Besides his work at FutureNow Inc., you will know Bryan from his many presentations and books. Most recently: Always be Testing.

What’s there to talk about here?

There are so many ways that web marketers can go after testing today: You can do it yourself, You can use free test automation solutions and of course high-end multivariate testing solutions.

There are ups and downs to each approach. Which approach is right for you and how do you get off the couch and into the trenches now?

And how do you get the time and $$$ that you need?

Even if you don’t use a commercial multivariate testing solution you still need to set aside time for experimentation. That time is money too.

To round out the panel, I will get to give a few minutes introduction and share the perspective on testing from the point of view of direct marketers. They have been working on it much longer than their web analyst colleagues, frankly.

As is the key theme in my Multichannel Marketing Metrics book, marketers from all disciplines have much to share with each other.

Don’t miss this rare opportunity. Join Josh, Bryan, and me next Thursday, Feb 5th at 12 pm US EST by registering today.


It’s a Predictable World! (Save 15% on the Predictive Analytics World conference)

These are great times for friends of business analytics. There are many telltale signs that after all these years, business analytics are still a rising star. Especially so, when it comes to advanced analytics such as predictive modeling.

Analytics, Superstar

In 2007 we were given the eye opening book Competing on Analytics by authors Davenport and Harris. While mostly a compilation of all the various kinds of analytics that have been used in the enterprise, its claim to fame is to motivate that analytics can be more than just rear-view mirror reporting.

Analytics can and should form the strategic basis on which companies compete.

The kind of math that Davenport and Harris find deserving of the name analytics are in fact predictive analytics, i.e. data mining and discovery of unexpected insights in data. In contrast, the kind of math used within most of today’s web analytics and business intelligence merely deserve the name reporting in Davenport and Harris’ description.

Enter: Predictive Analytics World, Feb 18-19 in San Francisco

For the very first time next month, conference chair Eric Siegel and team are bringing us the Predictive Analytics World conference. The timing couldn’t be better!

 Predictive Analytics World

Now more than ever, businesses require leadership from their analytics teams. We are all asked to do more with less. If predictive analytics are the top of the line, the kind of analytics that are most likely to form the strategic basis on which our companies can compete, then we all need to learn how, and we need to learn asasp.

In my Multichannel Marketing book I review

  • why it is so imperative for direct marketers to draw on predictive analytics for prioritizing their contact lists.
  • why brand marketers draw on marketing mix modeling to predict the effect of an extra marketing dollar spent on TV vs. radio vs. print, and even vs. online.
  • that web marketers can draw on predictive analytics for behavioral targeting
  • how central customer analytics teams at Wachovia (now merged into Wells Fargo) predict which of multiple promising offers should go out to each customer in order to make it most likely that overall results will be maximized.

But most importantly, I touch on practical challenges with getting predictive analytics right.

Why attend the Predictive World Conference?

I couldn’t imagine a better opportunity than the upcoming Predictive Analytics World conference for practitioners to learn first hand how to turn the theory into practice.

My personal prediction is that the math is the least difficult part of predictive analytics.

If nothing else, modeling software can take care of the math anyway and make it user friendly to Marketers.

What is much more critical however is to know how to apply the analytics for generating business value.

Of the many things you could analyze, where do you start and how do you go about it? Rather than spending 2009 doodling in the data, here is an opportunity to make predictive analytics a work horse for your company.

Are Predictive Analytics Worthwhile?

Companies that have competed successfully on predictive analytics include for example Capital One. They used to be a tiny, peripheral player and grew to be a dominant giant by getting predictive analytics right.

Take the survey

Predictive Analytics World starts being educational even before you attend. I encourage you to take the informal survey that they just put out. It doesn’t take more than 4 minutes but it will already teach you something, namely about more business applications for predictive analytics than you ever knew. I certainly learned something.

Plus, you can request to receive a copy of the survey results once the polls are closed. Then you will really know how your peers are using predictive analytics.

Get a 15% discount on Predictive Analytics World

The Predictive Analytics World organizers were kind enough to extend a 15% discount to readers of Multichannel Marketing Metrics. Use registration code


when you get your ticket and save a big chunk of money.

Not a bad start!

Fun with Web Analytics: Can You Measure Interaction With a Paperback Book?

Many marketers are still blissfully ignorant about web analytics and think that it was only of interest to the web team, namely for improving online marketing. Yet its promise goes much beyond.

This post aims to inspire you for taking web analytics up on this promise. 

Do you think web analytics could help measure and improve something as offline as a paperback book?

Book authors would love to know how deeply readers are engaged with the many pages that they have brought to paper. After all, people start reading many books but finish only few of them. Especially so with non-fiction books! reading

Having published a book on Multichannel Marketing Metrics earlier this year, I certainly was very curious how engaged my readers stayed through the 11 chapters that make up the book.Would they make it to the middle of the book and then lose interest?

With online content, this question is really easy to answer. For example, web analysts at the New York Times can easily tell for multi-page articles on, how many click through all the way to the last page of the article.

But how would you measure that for a paperback book?

Traditional offline marketers would be quick to recommend a solution based on something that they always use for measuring “offline marketing things”, namely the use of panels or focus groups.

You would go interview a few readers. Then you would extrapolate their answers to all readers in the hopes that your panel is representative 

But under the right set of circumstances, web analytics can help us do better than that.

Look for ways to feed your web analytics with offline data 

Throughout most of the 11 chapters of my book, I have provided convenience URLs (also called vanity URLs) that reference interesting additional information about the subject that is discussed on the page. There are some 25 shortcuts of this nature spread from the beginning to the end of the book, similar to the following (click to enlarge): 

Book excerpt with convenience URL

These convenience URLs redirect to the target destination where the information is actually to be found. Often those target URLs would be much longer and cumbersome to type in directly. Hence the name, convenience URL.

You see these convenience URLs used everywhere these days. For example, you will find them in the National Geographic magazine. You even see them blended in on TV programs.

The main purpose of these shortcuts is indeed to inspire readers with further information. But as a side effect, everybody following one of my convenience URL creates an anonymous ping on my web site. Web analytics can count all those pings and aggregate them up to a report. The following chart contains all the pings coming from the 11 chapters of my book where readers followed a shortcut: (click the picture to expand it)

How deep do readers read the multichannel marketing book?

What jumps out at you?

Looks like a nice bell curve of engagement that grows towards the middle of the book and then slowly subsides towards the last chapters. As readers get drawn into the middle of book they become even more likely to follow the shortcuts. Towards the end, readers become a little less likely.

This was great news for me, I am super pleased that the drop-off towards the last chapters isn’t higher. The last chapters of the book have been called the most powerful by some gurus in their reviews.

How can this help you? 

Much liked we used web analytics to measure engagement with a paperback book here,

  • Direct marketers can measure engagement with their emails, catalogs, and credit card offers.

  • Brand advertisers can measure engagement with their TV and radio commercials, or newspaper ads, or even outdoors ads in subway stations.

Now, don’t get too excited though.

You also need to know the limitations of this analysis. Let’s discuss a few typical ones in regards to the chart.

In the chart, you see a sharp drop-off within chapter 1 on the shortcuts that have been followed, although the 4th shortcut ( attracted more readers.

Why that drop off?

  • Is it because people didn’t finish chapter 1 and skipped to chapter 2? (Probably not, otherwise shortcut 4 wouldn’t have spiked up.)

  • Is it because more people are so familiar with shortcut 2 (/web20) and shorcut 3 (/brandweek) that they didn’t bother follow?

  • Maybe those particular pages were less engaging?

  • Or was it that the first two shortcuts led to content that wasn’t what readers expected so they stopped bothering at the third shortcut?

Two observations in response to these questions:

  • Web analytics can never really tell why people do what they do. That requires attitudinal data that is often best collected through surveys (ah, … so we are back to panels after all…)

  • The chart turns out not just to measure how far people get into the book. But the data is also influenced by how deeply engaging each shortcut was. 

    With those two aspects overlaid on the data however, does the positive interpretation that I originally made still hold true? Maybe fewer people made it to the end of the book but those who did were super engaged and eager to follow the hyperlinks because those really are the most powerful chapters?

 Welcome to the fun world of web analytics!

You see clearly why we need much more than just a web analytics solution to get good web analytics. We also need web analysts that spend enough time with the information to think about it, probe it, and verify their interpretations through experiments.

Alas, until the first edition of the Multichannel Marketing book is sold out and we get to publish a second edition, I won’t get to experiment with different shortcuts and re-measuring outcomes. That kind of thing would be much easier and faster with an e-book.


P.S.: If you are interested in hearing from some real life web analysts about their work, join the upcoming Web Analytics Strategies Panel moderated by Jupiter’s John Lovett . John will be joined by web analysts at The Hartford and Coastal Contacts.

View-Through in a Grocery Store ???

I love it when metrics worlds overlap! All the time, different marketing disciplines are coming up with comparable metrics but calling them different names, unaware of each other.

Did you catch MediaPost’s article that TNS Media has released a new offering for in-store metrics? Below is an excerpt where MediaPost captures the value prop:

“Dashboard combines information about where shoppers are in a grocery store at any given time, tracking the number of seconds they spend at any display, the amount of time they spend with other products, and then overlaying it with sales information. “A display’s stopping power is a good thing when it generates a lot of purchasing, but if people are spending many seconds there and not buying, something isn’t speaking to customers properly,” he says.”

Online marketers will dig this. What TNS is offering here is to calculate a view-through metric for in-store displays.

Now, TNS will answer the question: Based on how much viewing/attention/engagement the in-store display is achieving, what is the releative sales success for the products that it is advertising? If people are buying without paying much attention to a particular store isle, the display shouldn’t get as much credit, probably. If people however pick the product out of a special display area after studying it longer on average, the display probably should get more credit. Such displays should be used in other stores of the same chain.

I could not spot whether TNS will measure the interaction with the displays by putting people into the grocery store isle, evaluating cameras, or using a panel of volunteers and something like the portable people meter. Let me know if you have more info about this.

Go multi marketing discipline aware analysts!