Disorganized marketing data is as good as no data at all

As marketers, we play in big pools of data every day. We're renting lists, sorting and filtering databases of leads, and going through mounds of statistics to look for trends. That's great -- when the data is accurate, filtered and tagged properly, and the way it should be.

But do we take for granted that our data is good? This can be very dangerous. If you're renting a list and the list vendor hasn't done a good job of verifying the data and organizing it properly, you could end up reaching a lot of people you didn't want to reach.

Here's a great example. The other day I read a press release by Jigsaw, a site that's a cross between a list broker and a social network. (In case you're not familiar with Jigsaw, you can essentially buy, sell, and trade business contact information. It's billed as a great resource for sales professionals who need a constant stream of accurate leads.) Jigsaw's press release announced that it will begin providing all its company data -- company names, addresses, industry, number of employees, etc. -- for free. The hook, of course, is that you still need to pay for contact info for particular people within those companies. You can also license this data for use in your own applications, on your own website, etc.

This caught my eye because the media property I work for, IndustryWeek, is always looking for new ways to provide useful and relevant information to our manufacturing executive audience. I thought a licensing deal with Jigsaw might be worth looking into, if we could help connect manufacturing decision-makers with information about companies they need to know.

So I started digging deeper. I went to Jigsaw's Open Data Initiative pages, to a list of pre-defined filters they've already developed for the data. I was pleased to see a category and several sub-categories for manufacturing, which is exactly what I was looking for. So I decided to check it out. I clicked on the "Aerospace & Defense" sub-category, and the first few company names looked appealing. These all look like companies that are in the aerospace or defense business.

But then I scrolled down the page. I started noticing something disturbing. There were a lot of listings for martial arts studios. Wait a second: martial arts studios aren't manufacturers! It took a few seconds before I figured out what was happening. Their filter must be finding the word "defense" in the listings for karate studios, which of course is getting matched up with "aerospace and defense."

I thought maybe I just picked a fluke category, and most of the data here is good. So I tried the next sub-industry filter in the list, "Automobiles, Boats and Motor Vehicles." Again, the first few seemed good. But then I started seeing listings for "Funday Eco-Tours" and "Go Fish Charters." Again, these companies might be associated with boats, but they have nothing to do with manufacturing.

I'm saying this not to fault Jigsaw. I've played around with their site in the past, and I found it to be pretty good. The business model makes sense, although I can't say anything for the accuracy of the information because I've never checked it out personally. My larger point is that when you're given a data source, you should always ask questions like:

• How was this data collected?
• How were the filters/groupings/categories applied? For example, for a magazine circulation form, I'd ask if individual respondents put themselves into an industry grouping, if the publisher added the grouping based on the name of the company, or if it was done some other way.
• How fresh is the data?
• Was there any human intervention in the data collection process? In other words, did an expert review responses for plausibility? If it's numerical data, was it scrubbed to remove outliers?
  

As the world becomes more dependent on databases, we hear an increasing number of horror stories about data gone wrong (the TSA's "Do Not Fly" list comes to mind, where some people have been mistakenly placed on the list and go through a huge ordeal every time they need to board a plane). Don't let your marketing program become a victim of bad or improperly used data.

Buying advertising from ad networks versus individual websites: Which is the better approach?

Online advertising networks have been around for a long time, but they've recently seen a resurgence. An increasing number of website owners are selling their inventory through ad networks now, which I think is due to several factors. Targeting on ad networks has gotten better than it used to be, social media has increased the amount of ad inventory (particularly what I'll call "non-premium inventory" because ads on social sites typically have much lower click-through rates and engagement rates), and advertisers are more likely to pursue The Long Tail of their audience.

From an advertiser's perspective, should you be running your ads on a network or should you deal with individual sites for your buy?

First of all let me say it's a complex question that depends a lot on your particular targets, the types of sites in your market, and a number of other factors. But whenever possible, I believe advertisers will get the best value for their advertising dollars by dealing directly with sites that reach their target market, rather than buying inventory from an ad network.

This is true because most site owners know their audience well, and they can make recommendations for how to reach the audience more effectively. Sure, there are lots of site owners who aren't willing to help you, might not want to give you insight into their audience, etc. But I think the majority of online publishers realize it's a win-win-win when they work closely with advertisers. Their audience is better served by seeing ads that are more relevant, the advertisers are happy because they're getting better performance, and the publisher is able to keep its advertisers happy which leads to repeat business.

For example, at IndustryWeek we see both types of advertisers: The ones who truly want to partner with us on a campaign and really care about making their campaign fit the site's audience, and also the advertisers who are buying ad inventory and applying a one-size-fits-all approach. I'm sure you'd expect to see the former group's campaign results outperforming the latter. But I'll bet you'll be shocked when I tell you how dramatic the difference can be -- it's often 10 times better! That's right...if an advertiser does a good job customizing its buy, creative approach, and watching and optimizing the campaign, they can get 1,000 click-throughs on an ad when other advertisers might only be getting 100 clicks. It's all about being relevant to the site's audience.

A lot of media buyers -- whether they're buying for their own company, or working on behalf of an agency for a client -- say that online ad networks are important because of scale. They might say, "We'd need to buy from dozens or even hundreds of individual websites, and it would be so cumbersome that it's much easier to deal with an ad network." This is certainly true, and I'm sure there are lots of advertisers who run into this situation. In many of these cases, ad networks might be the only way to effectively deploy a large-scale ad buy. But before you opt for a network buy, think about that 10x return I mentioned in the paragraph above. If you're able to get ten times the return for maybe two, three, or five times the work, often it makes financial sense to opt for the more custom, more personal relationship directly with websites -- even if it means throwing more resources (time and people) at the campaign.

What's the best day and time for a webcast?

Here's an interesting question blogger Ken Molay is asking. What day of the week is best for a webcast? And what's the best time of day?

He's doing a survey to collect data, because as he notes, conducting tests on this sort of thing would be difficult. I completely agree that gathering hard data from tests would be tough, but I think most producers of frequent webcasts could get a pretty good hunch for which days and times pull well, just knowing some of the variables that are in play (like how popular the topic of each event is likely to be). On the survey front, I also worry about whether people's actions of actually attending webcasts at certain times will follow their responses, because quite often, what people *say* they'll do is quite different from what they actually do. But it should be an interesting survey nonetheless, and it's probably better than the hunches all webcast producers are operating on right now.

In July 2007, webcast vendor ON24 did an interesting study that looked at a lot of different data points with key webcast trends. They examined things like months of the year for webcasts, plus which days of the week are best for registration. But they didn't dive into the attendance question based on day of the week or time of day. (Here's a link to a press release about that ON24 report, with a contact name on how to request a copy of the PDF. I couldn't find a direct link to the report PDF anywhere, probably because ON24 wants your name first so they can sell you something. But hey, ON24 is the best webcast vendor out there as far as I'm concerned, so let them sell you!)

Here's my contribution to conventional wisdom on best times for a webcast...

Last October, IndustryWeek did 12 webcasts in two days. We called it the "Operational Excellence Online LIVE" web conference. Sessions started at 11am, noon, 1pm, 2pm, 3pm, and 4pm Eastern on each day (a Wednesday and Thursday). The target audience was manufacturing management. The majority of our audience is based in the Eastern and Central time zones, although there are still quite a few manufacturers in Mountain or Pacific. Overseas audience is generally around 5%.

Since there were lots of variables with these 12 sessions -- 12 different sets of speakers, 12 sponsors who did differing amounts of promotion for the event to their current customer/prospect base, 12 different topics of varying interest to the IndustryWeek audience -- it's hard to draw scientific conclusions. But in terms of attendee percentage, we can make some pretty good guesses.

We found that the first session (Wednesday at 11am ET) had the best attendance percentage. There were several others that had comparable numbers that were just shy of the first session's numbers, so I'm sure that's not statistically valid.

We also found that the 4pm session each day was by far the weakest in terms of attendee percentage compared to other sessions that day. This is despite the fact that both 4pm sessions had particularly good speakers and particularly interesting topics that are normally "hot button" sort of subjects for our audience. (We survey the audience after each session and ask them to rate each speaker, and out of 24 speakers for the 12 sessions, the speakers for the 4pm sessions were rated as the #1, #2, #4, and #8 presenters for the entire conference.) Good content, good speakers, poor turnout.

We had similar results for the 4pm sessions during our December 2006 multi-session online conference. Those two sessions underperformed from an attendance % standpoint as well.

As a result of these pretty convincing results, we've eliminated the 4pm time slot from future multi-session online conferences. We'll see how the numbers change in the future. Perhaps the 4pm time was too late for manufacturers, since a lot of manufacturing companies start and end their days earlier?

Webcasts: What's more important -- leads or thought leadership?

The April 7 issue of BtoB magazine has a graph with results from an online poll they conducted about webcasts. The question was "What's your top webinar objective?" and the two choices were leads or thought leadership. Interestingly enough, 69% said their top objective was thought leadership and only 31% picked leads.

These results really surprised me. I'm sure this wasn't a scientific poll, so perhaps I should take this data with a grain of salt. But in my many years of experience conducting webcasts, I'm accustomed to most sponsors being obsessed with leads. That's how sponsors usually evaluate success -- based on number of registrants and number of attendees, plus the quality of those people. It's just the way many companies are set up right now...they're dependent upon leads.

But after thinking about this poll, I realize just how difficult of a question "thought leadership or leads" is. A webcast needs to do both to be effective, so it's silly to say that one or the other is the primary goal.

If a webcast doesn't generate leads, it's going to be difficult for the sponsor to quantify results. How will sales and marketing turn that thought leadership into action without leads? But on the other hand, if the webcast doesn't provide thought leadership, the leads are actionable but they're not nearly as valuable. After all, there are a lot of easier and less expensive ways to get leads. But the thought leadership conveyed during a well-executed webcast, combined with the fact that the person just sat through an hour-long presentation about a topic, makes for a very qualified lead.

This is exactly why I believe webcasts are so popular in B2B markets right now. It's not about the lead -- nor is it about the thought leadership. It's an inseparable combination of the two.

How much time do you spend fighting fires each day?

Today was a jam-packed day at the IW Best Plants Conference. (See yesterday's post for more). I heard a number of interesting presentations about how manufacturing companies can implement continuous improvement in their facilities. A lot of the challenges and solutions discussed today were of course specific to manufacturing, but there are some things that digital marketers can learn from too -- no matter what industry you're in.

During a session called "How to Create a Continuous Improvement Culture," David Rowland from Milliken & Company (one of the world's largest textile companies) talked about "firefighting" and how it's often a significant portion of an employee's day. How many "fires" are you putting out each day and how many crises are you solving, versus doing your normal job?

David said that the typical person in a typical company spends about 40% of their time on daily operations (the "normal" part of their job), and 60% trying to solve unexpected problems and put out various "fires." Milliken & Co. actually did a study a while back where it examined how its managers were using their time, and the numbers were amazingly close to this 40%/60% ratio.

However, he pointed out that the best companies -- like Toyota -- only spend 20% of their time on daily operations and 20% firefighting. So where do they spend the other 60%? Continuous improvement. They make their processes better. They find ways to reduce waste. They standardize routine decisions.

So how does this apply to online marketing? Well, just like the people in manufacturing operations, we can put processes into place that eliminate variability and waste. Whether it's a report your team puts together for management, a process for distributing leads to sales, etc., implement a plan to standardize these types of activities. As David said during his presentation, "People love their jobs when they need to make fewer decisions. They don't need to make many decisions when you have good processes in place."

But here's the tough part: no process is ever finished. Once you have the process in place, always look for ways to make it better, like Toyota does with its famed Toyota Production System in the manufacturing world.

As digital marketers, there are plenty of things we do that can't be standardized. But I think we underestimate the number of standard tasks we complete each day -- and you can apply continuous improvement principles to most of these. Standard practices will reduce the amount of time you spend "firefighting" and give you more time for the big, difficult, custom tasks you perform.

What marketers can learn by looking at a manufacturing plant

I'm at the IndustryWeek Best Plants Conference this week. More than 600 manufacturing leaders have gathered in Milwaukee to learn best practices from experts and their peers. Plant tours are one of the key components of the conference, so manufacturers can see other plants' operations and learn from them. Today I was the IW representative who captained the Miller brewing plant tour.

How much can an eMarketer learn from a manufacturing operation like Miller's? Two things came to mind as I went through the tour:

1. Digital marketers should follow the example of manufacturing companies that post their key metrics and list of improvement initiatives for all to see. On the Miller tour today, we went into the room where teams gather for 20 minutes before each shift starts. The front of the room is covered with at least three dozen graphs -- colorful, filled in with magic marker -- that show their key performance indicators and how each shift performed over the past month. They post info about productivity, safety, attendance, etc. There's also a couple whiteboards that enumerate specific tasks they need to undertake, maintenance issues that need to be addressed, and problems that need to be solved.
   
   We need to do the same thing as marketers. Too often we keep our numbers buried in a spreadsheet. Even if your metrics report gets emailed to the relevant people and they read it today, the numbers are quickly forgotten tomorrow. Try posting a board in a prominent place in your office that shows your key metrics and how you're performing against them. Reports and spreadsheets that aren't posted don't cut it...because they're too easy to hide and too easy to forget about.
   
   
2. You need to have a passion for what you do. We didn't get to see a ton of employees -- maybe a dozen over the course of the three and a half hour tour. But the employees we saw were obviously passionate about the 150 year tradition of Miller and its mission to produce great beer efficiently. I could see it in their eyes and I could hear it in their voices.
   
   When's the last time you felt passionate about an electronic marketing campaign? Sometimes we get caught up in the daily grind -- churning out one project after another -- that we forget what it's all about. I know I'm often guilty of that.
   
   Dig a little deeper. Make the extra effort. Spend the time to do a great job on the little details. Remember what attracted you to this business in the first place. And once you've finished -- when you've produced a campaign or materials that are really first rate -- kick back and crack open a cold one...just like they'd do at Miller.
   

Keys to a successful campaign

As a media provider, my team is often asked to answer marketers' questions like, "What's your average click-through rate on [X web ad placement]?" and "How many leads do you think we'll get if we run an ad in [Y newsletter]?" When I put myself in these marketers' shoes, I completely understand what they're trying to find out. These are good questions to ask.

But probably like many other media providers, we don't answer these questions with exact numbers. Sure, we're happy to provide a range so the marketer can get a feel for how well their ad might perform. But usually we'll present it in the context of "the best ads get X percent" (or X number of conversions) and "the worst ads get Y percent" (or Y number of conversions).

The reason we offer a high/low range is simple. In most situations, either the marketer or agency is controlling the creative message for the ads. During the buying process, when they're asking these types of questions, we have no idea how good (or bad) the creative will be. Well-designed, well-written, and well-thought-out campaigns will get much better results -- independent of the medium.

Even though it covers a lot of points all of us have heard before, this article by Harry Gold of ClickZ does a good job of reminding us of the keys to a successful creative execution. Start with the message and offer, then think about the landing page, and follow through to the post-action phase like confirmation pages and confirmation emails. Every step matters. Good ad creative can't overcome a bad landing page. A good landing page is useless if the creative message is wrong. And if a strong follow-up isn't in place, you're throwing away perfectly good opportunities.

One thing I'd like to add to Harry's article: Don't forget about what comes after these steps too. How will you nurture these leads? What's your sales team's approach for following up, taking the prospect through the sales cycle, and closing the sale? Often these tasks are not our job as marketers -- since this responsibility is passed along to sales in many companies. But if there's not a seamless plan in place, you know what happens. Leads and opportunities are lost, even if you did your job developing a successful advertising or marketing campaign.

Google Analytics offering benchmarking data

About two weeks ago, Google announced that its popular Google Analytics service will now offer web stat benchmarks. According to this post in the Google Analytics blog, the benchmarking tool "enables customers to see how their site data compares to sites in any available industry vertical." So you'll be able to compare your web traffic patterns against others in your broad industry grouping.

Of course no site-identifiable information will be shared, and the minimum benchmarking pool will be 100 sites. You won't be able to view info on a particular competitor, but rather an aggregate group of 100+ comparable sites. In addition, Google will group sites based on their relative size (small, medium, or large), so you won't be comparing Mom & Pop Inc. with Microsoft, even if both are software companies.

This is a great idea, and something I'm not surprised to see Google offering. They have the data, so why not provide it to their users? In the long term, helping sites optimize their web traffic can only benefit Google. Presumably a site that's watching traffic patterns more closely and optimizing its performance will be a better Google AdWords customer when buying pay-per-click ads, so this seems like the next natural progression.

Benchmarking data is likely to be a double-edged sword for webmasters and web marketing managers. While it's great to compare yourself against an industry average when you're doing well, with these new benchmarks I have a feeling a lot of web pros will find themselves explaining sub-par performance to senior management more frequently. It used to be a lot easier to say "we're doing well" when the monthly stats report is passed around the company or department, when there was no data to benchmark your performance!

But the smart marketer will welcome this information. As the old adage goes, you can't improve what you don't measure.

Let's replace the term "open rate"

Many people have discussed the problems inherent in the "open rate" statistic you might see in your email marketing reports. Not to get into the details again, but basically it's a highly misleading number -- it doesn't truly give you a good count on the number of messages that are open for a variety of reasons. With Outlook 2007, Gmail, Yahoo Mail, and others blocking images by default, chances are good you've seen your "open rate" stat drop a lot in the past year or so.

As an industry, let's rename "open rate" to something that implies the inherent problem in the numbers. The metric really doesn't measure the number of people who opened your message -- it measures the number of people *we can track* who opened it. There are probably lots more who actually opened your email that aren't accounted for in this number.

I propose "counted open rate" as the new name for this number, because that is really what the metric tells us. It shows us the number of people *we were able to count* who opened your message.

A subtle difference? Perhaps. But I believe the word "counted" at the front of the metric may remind marketers this isn't a perfect number. It might lead senior execs who don't understand the nuts and bolts of digital marketing to ask, "What does 'counted' mean, and who are the uncounted opens?", which of course sets up the discussion about the metric's inherent flaws.