With all the buzz around social media marketing, more and more companies are jumping on the social media bandwagon. Social media offers a greater reach than some of the traditional online marketing channels, and enjoys fewer competition. Yet, there’s a lot of skepticism among some around social media marketing. The main roadblock that we’ve seen so far has been around measurement. Many companies still have the perception that social media is not measurable or that it can not be measured with the same standards as other online media. The following is a list of myths and facts about social media measurement.

Myth: I don’t pay for social media traffic. Since I don’t pay for it, I don’t need to measure it.

Fact: This is only a perception. The fact that you’re spending your resources – time and labor – means that you’re paying for social media traffic. True, there’s no cost-per-click model, but the mere fact that you’re spending time writing blogs, creating Facebook pages, responding to people’s requests or questions on Twitter means that you’re spending valuable resources on social media. Like any other activity, there’s an opportunity cost associated with your efforts – time you could be spending doing other things.

For example, if an employee costing the company $80,000 per year spends 10% of his/her time responding to blogs, Twitter posts, etc., then the cost of social media associated with that employee alone is $8,000 per year.

Myth: There’s no measurement standard when it comes to social media measurement.

Fact: There has been no shortage of measurements introduced when it comes to social media marketing – video views, sentiment metrics, advertising equivalency, etc. What’s important to note is that the same standards that have applied to PR measurement apply to social media as well. For years, PR professionals have classified the measurement of their activities in one of three buckets:

• Outputs (who’s talking about you): this is the basic level of measurement needed. Mapping this to social media measurement, the metrics associated with this pillar are video views, number of RSS subscriptions, number of blog or news mentions, etc.
• Outtake (what are they saying about you): this is the qualitative side of PR and social media measurement. Many brand managers use social media measurement for this aspect. The metrics associated with this pillar include qualitative measures such as “thumbs up”, “thumbs down”, measuring influence on Twitter or within the blogosphere, etc.
• Outcome (what does it mean to your business): this is the ROI portion of the measurement. How much site traffic and conversions can be attributed to social media. Do visitors who viewed a YouTube video end up visiting the site? This is clearly the most critical aspect of social media measurement for online marketing professionals.

The same standards that have applied to PR measurement also apply to social media measurement.

Myth: You can’t measure the ROI of social media.

Fact: This is not true. This is exactly what Tealium Social Media has been designed to do. Whether you measure your results based on the amount of traffic generated on your site, number of leads generated and the online sales volume, social media can be measured accordingly. In fact, so far most of the interest from customers has been around using measurement to audit their social media and PR agencies. Yes, you can measure the ROI of your social media marketing.

Myth: I can’t compare social media to other online marketing channels.

Fact: Again, not true. By integrating data from Tealium Social Media into your web analytics tool you can measure the ROI of social media side by side with your other online marketing channels such as search, email and banner advertising. This means that you can in fact compare your social media traffic – traffic attributed to video sharing sites, bloggers and news coverages – to your paid marketing channels.

Yes, social media is measurable and because it’s still a new practice, this means less competition and therefore better ROI than other online channels. We are working on a number of case studies in this area and are looking forward to sharing our findings once completed. Stay tuned.

I was watching the wild card playoff games between the Falcons and Cardinals where I noticed a huge media buy by Hyundai for the Hyundai Assurance program. According to the official press release,”Hyundai is providing a complimentary vehicle return program for the first year on every new Hyundai that is financed or leased for owners who experience an involuntary loss of income within 12 months of the purchase date”.

Certainly this is an unprecedented move in the auto industry. And marketing wise, it is a brilliant one. According to the American Marketing Association, marketing is defined as “the activity, set of institutions, and processes for creating, communicating, delivering, and exchanging offerings that have value for customers”. This program certainly provides value to customers in these economic times.

And after only two days, it looks like the program is so far getting a great deal of coverage through the press, blog community and social networks. Congratulations to the marketing folks at Hyundai for creating a relevant marketing program for the times.

A while back we published a post on using Google to predict elections. A similar post has recently been submitted by Jeremiah Owyang on the use of social networking stats for similar analysis. Obviously, in both cases, one cannot heavily rely on these numbers to predict elections, as they’re a reflection of interest and one cannot guarantee that interest turns into votes. But in both cases, there’s a great deal of entertainment value associated with the data.

However, a recent article in New York Times reveals a much more powerful side of Google Trends. In this case, Google is using its search data ot track the spread of flu within the United States. More specifically, deta from last year revealed that Google saw a spike in number of searches for terms such as “flu symptoms” about two weeks before the CDC (Centers for Disease Control) reported regional outbreaks. The graph below shows the comparison between the two data sources.

I must admit this is one of the most clever uses of Google Trends as it captures real data in a manner that can be used to actually predict trends. Congratulations to the folks at Google.

I always enjoy what the guys at Common Craft do. They have a way of simplifying things for everyone to understand. This is another fantastic piece of work by them. It explains social bookmarking. Many people miss the social aspect of bookmarking.

[vimeo 267072]

One of the most popular KPIs for lead generation sites is “Cost per Lead”. It lets marketers know whether they’re spending the right amount on marketing campaigns. However, a better KPI for such customers is “Cost Per Qualified Lead”. It provides a more accurate picture of the campaign performance.

The following is what a customer has recently shared with us regarding their use of WebToCRM.

Note: the numbers have been revised in this post.

After running a number of online campaigns, they were able to use their web analytics solution to determine high-level performance metrics such as campaign clickthroughs, conversion rates and cost per lead. The sample data is shown in the first graph below.

The campaign conversion rates are mostly within the same range. However, because of the higher conversion rates and lower cost per click (CPC), newsletters end up generating a much better cost per lead (CPL). Based on this data, it would make sense to take some budget away from search engines, which have the highest cost per lead and put that money towards more newsletter sponsorships.

However, the customer did not stop there. Thanks to WebToCRM, they integrated their online campaign data into their salesforce.com implementation and decided to break down their leads into two classifications:

• Qualified leads: these are defined as leads with complete and accurate information, including the person’s full contact information, job title and decision making timeline.
• Unqualified leads: these are leads with incomplete or inaccurate information such as fake email addresses, etc.

How do the numbers hold up when you look at qualified leads instead of just the raw number of leads? The picture turned out to be quite different and is shown below.

With a simple “Cost per Lead” KPI model which is what many web analytics practitioners use, the company would have diverted money from search engines into newsletters. However, the more relevant “Cost Per Qualified Lead” KPI shows that the customer would be well served doing the exact opposite. Although search engines provided fewer leads per click than newsletters, they also provided a far higher percentage of qualified leads. The client is therefore going to continue its investment in search engine marketing.

Still wondering about the value of cross-channel analytics? Think it’s expensive? Take another look at WebToCRM. It lets you integrate your online campaign data into your CRM application, regardless of what web analytics or CRM tool you use. Best of all, the Free edition give you the same level of data that you see in this example.

Ever wonder what all the buzz is about social media? The folks at Common Craft have done the best job I’ve seen explaining social media. Enjoy.

[vimeo 1083838]

Media for the people, by the people.

Ever tried measuring the return on investment of your press releases? Although seemingly simple, the process turns out to be more complex than one might think.

Recently we published a press release announcing the WebToCRM product. For those of you interested, it can be found here.

The purpose of the press release was to introduce WebToCRM, but also to generate traffic and leads to the tealium.com Web site. So how did the press release do? It turned out to be quite successful as shown in the chart below. It ended up generating a tremendous boost in traffic.

The site traffic on May 20 increased exponentially on the day of the PR launch. Given the fact that Tealium is a startup and the site does not get much traffic today, clearly the increase can be attributed to the press release.

But drilling down into the data reveals a hidden value of PR that cannot be detected by web analytics solutions. To analyze this more, we’ve broken down the traffic by source. The categories analyzed are “Bookmarks or Directly Referred” (typically direct traffic to the site or emails), “Search Engines”, “PR Sources” (compilation of referring URLs associated with the news stories), and others (not fitting any of these mentioned categories). The results are shown in the figure below.

We’re analyzing two seven-day periods, one before the release and the other after the release. The results are definitely revealing. Obviously, the PR traffic increased since there was no PR the previous week. But more interesting is the fact that both the bookmark and search engine traffic increased dramatically as well. In other words, there’s a hidden value to PR that previously hasn’t been mentioned in the Web analytics world. Because of the press release, more people came to the site directly and more people searched for “Tealium” on search engines.

If you’re a marketing veteran, the results make sense. Press releases are a great medium for generating awareness. Also, by creating a marketing mix that includes multiple touch points (PR, search, banners, trade shows, etc.), you’ll get an overall result that exceeds each one of the touch points combined. In other words, 1+1+1=4.

However, it also reveals the challenges associated with PR measurement. It shows both the effectiveness of online PR, as well as the difficulty of measuring it. In our case, this was not a difficult task, because of the reasonably low traffic that we were getting before the press release. But if you’re a high traffic site, measuring the effectiveness of PR becomes much more challenging and certainly an area that has not been addressed by Web analytics solutions.

The field of PR measurement today consists of understanding your Outputs, Outtakes and Outcomes. Outputs and Outtakes are fairly simply to measure, but Outcomes (the business results of the PR – site visits and conversions in the online world) are much harder to measure, and ironically more valuable. We’d love to hear back from you about your experiences in this field and what you’ve done in your organizations to solve this problem.

An unscientific approach!!!

One of my favorite tools to see trends, patterns and seasonality associated with search terms is Google Trends. It lets you see trends associated with specific keywords and compare up to 5 keywords together.

With all the buzz around the democratic primaries, it was only fitting to use Google Trends to see if we could see patterns that could shed some light into the outcome. The results, although not scientific are pretty revealing. First off, here’s a comparison of searches for the terms “barack obama” and “hillary clinton” over the last 12 months. We’ve obviously filtered out the international traffic and the results are shown in the figure below.

The interesting trend here is that Obama was behind Clinton in terms of searches till January of 2008. January 3rd happened to be the date of the Iowa caucuses which showed a surprising win by Obama and one can speculate put him on the map as far as the general audience is concerned. To test the hypothesis, lets look at a similar comparison, but this time between “huckabee” and “mccain” and interestingly, we see a similar pattern.

Within the GOP front, we see a spike in interest for Huckabee prior to the Iowa caucuses and a decrease after McCain gains momentum from consequent wins.

How accurate is this?

So you’re wondering, how accurate is this? While Google Trends is a great tool for search engine marketing, it is simply not built to forecast elections and markets. For example, if you look at the breakdown by states, you can see that the term “barack obama” gets higher traffic than “hillary clinton” in all states, including Pennsylvania, Kentucky and Ohio – where Hillary Clinton won by a large margin. This is shown in the image below.

In fact, in terms of popular vote, both candidates were neck and neck. And you can make the argument that most people did not know much about Obama before the primaries began and therefore what we’re seeing is people educating themselves about the candidates. But it’s certainly fun to see the trends in terms of peaks and valleys around some specific events such as the start of primaries and caucuses.

The purpose of this post is not to endorse any one candidate or make market predictions, but rather showcase what you can get from Google Trends. The information can be very useful in determining peaks and valleys in user interest associated with specific events of relevance to your business.

In the traditional offline media (TV, radio, print), agencies typically charge 15% of the marketing spend as their fees. So for example, if your company is spending $100,000 on print advertising, the agency that you’ve hired gets $15,000 for their services, which could include ad design and creation, media engagement, etc.

I recently came across a great column by Kevin Lee mentioning that because of all the optimization work that has to be done around SEM, the 15% fee may not be enough and clients should be willing to spend more on agency fees. Kevin is right. There’s a lot of work that needs to be done in the area of SEM optimization. Here’s a few must-dos for an effective campaign:

• Research the list of potential keywords to be targeted
• Select the keywords to be targeted through SEO, those through SEM, and those through both
• Create the right ads for keywords. You simply cannot use the same ad for all your keywords and the rule applies as you go from one outlet to next
• Determine the budget to be allocated to each
• Create specialized landing pages for each ad
• Well, you get the idea …

This becomes especially an issue with smaller clients since they won’t be able to leverage the economies of scale and scope enjoyed by larger clients who spend more money on more keywords.

At the same time, I couldn’t help but notice the disincentive between CPC and agency fees. I understand why agencies use a 15% fee on offline ads – there’s no viable measurement beyond the output. You can only reasonably measure the number of times and the money spent on ads. But in the online world, things are different. You can measure the entire visitor cycle – from impressions, to clicks to conversions.

Think about it: would you rather spend $10,000 and generate 10 leads or spend $1,000 and generate 50 leads? Of course you’d like the latter, but in the world where agencies are paid by CPC, their incentive is to spend more money. They’ll benefit from having you spend $10,000, whether the campaigns are optimized or not, whether the conversion funnels are streamlined or not.

Ideally, agency incentives should be the same as the clients: get the most you possibly can from your marketing budget. This means that their fees should be tied to your conversions – online purchase, leads, registrations or whatever they may be. Don’t get me wrong. I’m not saying agencies should not be paid for all the optimization work that they do. They absolutely should. But perhaps it’s time to introduce more complex pricing schemes that do every body justice: a one-time fee and maintenance fee to cover the work to be done to get the campaign going, and a conversion/engagement bonus that gives the agencies the incentive to outperform.

This way, the agencies will get compensated with all the leg work necessary to get started and will be incentivized to constantly outperform, to constantly work on ways to increase their campaign conversions.

I’d also like to hear about some of the payment schemes you’ve used that have worked for you.