Newmarketmind.com

You read it on every blog, in every marketing book, at every industry conference: Segment! Segment! Segment! But what to do when you have no budget and very little time?
To add to this daunting mission, you discovered that your company legacy systems are disparate and they don’t talk to each other. Each department holding a crucial piece of the customer data like Milton was holding his stapler in “Office Space”.

Before you run to the hills and find another job selling granola to sheep shavers, there are concrete steps that you can take right now on the path toward full digital intelligence.

All you need is planning and some smart implementation decisions using low cost or free tools

5 steps to get started with segmentation:

1. What campaigns do I want to launch?
2. What information do I need to know about our customers?
3. How can I collect this information?
4. How do I deliver my campaign to a specific customer segment?
5. How do I Optimize my campaigns

Step 1: define an initial set of campaigns:
• Map out the customer cycle on your site
• Brainstorm for opportunities at every stage of this cycle
• Draft campaign ideas and messages

Step 2: What visitor data do you need to support your campaigns?
• Go over each of your drafts and find the underlying data you need to properly execute them.
• Identify data sources
• Compare the set of data you need with the data you already have in your systems. Identify gaps.
• A single visitor identification is important, especially if you are merging new with old data.

Step 3: design and implement a data collection and consolidation strategy
• Determine where the data will be stored and aggregated
• Tag your digital properties to collect the required data
• Use Tag management solutions to abstract the data collection logic from your legacy systems

Step 4: implement the campaign delivery processes to your customers segments

• Create specific business rules,
• Evaluate message delivery options: email, on-site, off-site.
• Create message templates and content
• Launch your campaigns from easiest to execute to more difficult

Step 5: Optimization
• Harvest customer’s emails early in their buying cycle in order to have more opportunities for retargeting (off site) or remarketing (email, on-site)
• Implement “lazy registration” design pattern to collect customer behavior even when they choose not to register
• Test various messages to find the winning mix of offers and/or discounts

Frederick Buhr

No, it’s not the smoking, the boozing or even the womanizing… Google is just introducing  new metrics  to measure online advertising.

In a nutshell,  ”active view”  will tell us if an ad has been watched or not and “active GRP” will measure how many times the right message is hitting the right audience.

So what’s the big deal? Well, this rather dry and technical announcement is big news in the microcosm of online advertising. Just check for “Google new metrics” on Twitter…

A lot of professionals think it is a move in the right direction, others have more questions to ask before they decide if this is good or bad (the school of testing of which I am an evergreen student). A third group, more cynical, think that Google is gunning for TV ad money.

Obviously,  the old ” impression and click” model is not enough to sustain Google growth.

Wall street analysts and big ad agencies will probably embrace the news.  Next thing you know, some large TV ad budgets will be allocated to online advertising and the giant of Mountain View will continue its march towards more revenue and a bigger valuation.

But there is a strange twist to this story: when Google introduced Adwords in 2002, it was earth-shattering  for the advertising industry. Many Madison Avenue agencies scrambled to adapt, those who did not saw their clients leave  to go to smaller interactive shops.

Google turned the world of traditional advertising upside down ; clients could now track the impact of their campaign with a precision that would have made a Swiss watchmaker proud. The old adage about not knowing which part of your budget was responsible for the sales was thrown to the dogs.

10 years later, Google seems to recant on the revolution it ignited. The geeks that changed the world have a new model: Madison Avenue!  They want to sit in Don Draper office, smoke his Lucky Strike, drink his Johnny Walker and win big accounts by being suave and macho.

Is Mad Men really inspiring Google? Is the highly trackable success of an ad campaign a thing of the past? Do we now need to measure ad campaign like the 60′s to break today’s glass ceiling?

If these Madison Avenue metrics give us an extra edge in planning and managing online campaigns they  will be a welcome contribution to the revolution.

If they are blurring our vision by  introducing fuzzy reporting and make us spend more money, I will be mourning the end of an era.

Frederick Buhr

At NMM we like to manage SEM campaigns manually, staying away from rule based bidding tools because they rarely spot trends. Only brains+data spot trends.

To spot trends that will help us manage keywords inventory, tweak ad copies and revise landing pages (and many other things) we first look at entropy.

Frederick Buhr, founder of New Market Mind.

In information theory “entropy” measures “ the quality of information carried by any signal”.

Applied to SEM, entropy can be determined by the amount of page viewed (information) carried by any visit (input signal) necessary to perform a business action (output signal) within the current environment of your web site . Of course, other type of information/input/output combinations can determined entropy but this one is simple and meaningful at the same time, so let’s start with it.

the Input Signal (IS) is the CTR of your ad (example: low IS = 1%, medium IS is between 1 and 3% and high IS above 4%).
This signal will measure the relevance of your ad copy to the keyword or search phrase you are bidding on. It needs to be as clear as possible, only [exact match] keywords or phrases should be applied to calculate IS.

Page viewed (PV) is relative to your site average page viewed per visits as reported by your web analytics. Low In is below the average page viewed, Medium In is around and High In is above.

Output signal (OS) is relative to the average conversion rate of your site as reported by your web analytics. Low OS is below the average conversion rate, Medium OS is around, and High OS is above.

Any score will deliver interesting insights but some combinations are like pointing a laser beam inside a dark tunnel:

Low IS – Low PV – Low OS = Keyword not relevant to your ad, your landing page and your offer. Stop bidding.
Low IS – Low PV – Medium OS = Don’t give up! introduce more ad copy, add negative kw, change landing page or improve the actual one. This combination could be a sign of seasonality or an emerging trend.
Low IS – Low PV- High OS = Increase bidding, try more ads, add kw variations, add negative kw.
Low IS – Medium PV – Low OS = kw used by lookers. Could also be a landing page problem. Do not bid on top position unless it’s a bargain.
Low IS – High PV – Medium OS = kw used mainly by lookers. Reduce bid but stay with top 3 positions (only if you can afford it) Landing page should be optimized for organic ranking on that kw.
Low IS – High PV – Hight PV = Buyers! Raise the bids and get all the traffic you can get at the right cost. Optimize for organic ranking.

Medium IS – Low PV – Low OS = kw probably not relevant to the offer or Ad copy/landing page combination is misleading.
Medium IS – Low PV – High OS = kw used by impulse buyers. Get more traffic if your offer remains competitive, use a coupon.
Medium IS – Medium PV – Low OS = kw is relevant, need the right offer on the landing page. use for re-marketing.
Medium IS – High PV – Low OS = KW used by lookers. Optimize for SEO. Add kw to remarketing campaign.
Medium IS – High PV – High OS = Buyers kw. Get more traffic by raising bids and add kw variations but watch cost. Keep an eye on competitive ads.

High IS – Low PV – Medium OS = Are you overspending? Lower bid, use coupon/offer on landing page to see if conversion goes up.
High IS – Low PV – High OS = Good combination of keyword, ad copy and landing page. Bid top position but watch rising costs.
High IS – Medium PV – Low OS = lower bid, check ad copy could be misleading or landing page confusing.
High IS – Medium PV – High OS = raise your bids to grab more traffic, watch the competition. be ready for a bidding war.
High IS – High PV – High OS = congratulations you have reached SEM nirvana!

This is a basic entropy model, you can always add more signals that are relevant to your business.
Bidding tools have very sophisticated logic, they are very expensive and they act on patterns. Entropy models are easy to set up, monitor and interpret. Best of all they will expose audience behavior that could become new business trends.

Thank you for reading!

Frederick Buhr

Our customers are great and we learn so much from them! Looking at their web analytics reports (under a signed NDA of course) we do not see a significant influence of the “share engines”: Facebook, Twitter or Youtube in delivering traffic to their web sites.
However when we analyze their social media presence we see amazing things.
In some cases, 50% of wall interactions on FB are with customers and are related to support or service.
More precisely, in the case of a travel site, 30 % of interactions came from planners asking about a destination, 30% from shoppers with questions about specific products and 40% from past customers asking for support while traveling (this one blew us away quiet frankly)!

Based on this data, we are assuming that most of our customers websites are driving traffic to their own Social Media page…
Should’t it be the contrary?

Investing in social media could have a different meaning, instead of spending marketing money on tiny ads, silly contests and funny videos to lure people in and build a community around a brand, companies could invest in giving better support to their customers enquiries.

Historically, for most companies, these inquiries used to only come through the phone or via email and be handled by the Contact Center and customer support teams.

The rise of social media has changed all this and different resources are now involved in answering product questions, helping customers get the full benefit out of a product or service, troubleshooting or assisting customers to exchange or cancel an order.

We were surprised that most requests coming from FB and Twitter required a high level of product knowledge and customer service expertise.
In addition, we have found that responding to customer service inquiries through the main corporate Facebook page or Twitter accounts could devalue the appeal of those marketing channels by mixing complaints and rants with the happy stories of marketing.

As a consequence, we have advised our customers to implement the following initiatives:

• Open a new Twitter account solely dedicated to answering customer service inquiries and product questions
• The main Twitter account will only focus on building brand awareness and broadcasting marketing messages
• The Marketing team should focus its efforts on managing the main corporate Twitter account
• The customer service Twitter account will be managed by expert resources within one of the company’s customer service groups (call center supervisors are ideal candidates)
• Customer support questions will be handled by expert resources within customer service group.

This organization is not meant to shift sales from one channel to the other. Customers need more than a quick wall post before buying and this information needs constant updating. Corporate web sites and call centers are still best fitted to contain and maintain that type of information which have also legal ramifications.

Regardless if your company does a good job in the market place, its reputation will be carried over into social media. The way you will support your customers on the “share engines” will make all the difference and could even make up for the “perceived” shortcomings of your products and services.

Frederick Buhr

I have been reading a lot of strange things about social media marketing lately.
For one, that campaigns are not directly trackable, as if you cannot directly link sales to a twitter promo or from a Facebook banner.

Some social media marketers think it is best to declare this medium only measurable
the old fashioned way — by establishing a baseline of traffic and sales with before and after results. Some even suggest taking the old FRY (Frequency, Reach, Yield) measurement out of the broom closets of Madison Avenue.

The major problem with this antiquated concept is that it could only work in a vacuum the size of the electron-positron collider. What about all the other marketing channels? Do we need to stop utilizing them while measuring the effect of social media so their (large) impact does not pollute our experience?

This solution is not only ludicrous, but it could undermine the foundation of online marketing as the most trackable sales channel ever created.

Do we want to be thrown back into the dark age of marketing? When print, TV and radio ad campaigns sucked billions of dollars from companies without delivering any concrete proof of direct impact on sales?

In 2000, search engines and web analytics began to rock that opaque advertising model, paving the way for Google and its keyword-based advertising model featuring pay-per-click, real-time bidding and tracking. Marketing budgets of all sizes moved away from obscurity and Google became the largest advertising medium ever.

This migration had a profound impact on the believers of FRY. Many disappeared, others found themselves on the verge of bankruptcy, only avoiding it by reducing staff and implementing dramatic cut backs. Today, faltering offliine channels are mostly relegated to the role of widening the online conversion funnel. Thus, making more customers Google a product or click on a banner. Here we have an evolutionary tale with an ironic twist: the old order is not dead yet. It’s feeding the new and helping it grow.

Who would have imagined that digital marketing could grow an obscure side? It is in danger of doing so if we, responsible digital marketers, fail to measure Social Media the same way we are measuring SEM, affiliate marketing and email campaigns. Which is consistently, accurately and using the same analytic toolkit loaded with relevant attribution models per channel.

Too bad if the ROI stinks! Who is expecting a banner on Salon.com to deliver sales? It should generates leads.
By the same token, we are not breathing down an affiliate’s neck to deliver branding. Just sales…at the lowest possible cost.

Social media marketers should not stray away from the virtuous practice of direct tracking and measurement unless they are ready to introduce a dose of charlatanism. Adding voodoo analytics to the social media hype will not make it the new SEM — it will only precipitate its downfall.