The main question is (as it was then) what a decrease in consumer spending and the knock on effects of that mean in reality.

First I looked at what happened the last time the recession came around and what caused it.

The main reason was (as now) the savings and loans in the US began to collapse, meaning that the institutions providing finance for mortgages were too generous to the public resulting in over borrowing. Over borrowing led to late and non-payment of debts, resulting in property repossessions. Repossessions meant a decline in housing demand. A decline in housing demand meant the value of property dropped drastically meaning that the loans secured against mortgages meant that many house owners went into a negative equity situation, IE; the mortgage costs more than the value of the property.

Because of negative equity institutions didn’t allow any further borrowing against mortgages. This means that typically high ticket items like cars, boats & property improvements were the first kinds of business to suffer from the situation. It also means the suppliers of these industries suffer.

As the knock on effect got worse businesses start to lose jobs and suddenly instead of employment improving and companies growing we see a recession. Add the gulf war of the early 90’s and increasing fuel prices and it all sounds a bit gloomy.

When the US had problems then the rest of the world followed suit due to a weak dollar (that we’re all tied too), poor international trade and therefore similar credit problems appearing all over the world.

The similarities with the 1990’s and now is uncanny.

Based on what I found out my view is quite simple. The recession means that weak operations will go out of business.

Weak operations

Businesses that have a middle of the road brand, no effective measurement, high prices without a high value proposition or businesses with high risk purchase options - such as hire purchase or credit financing on high priced consumer goods. Luxuries will be cut back on, but they won’t suffer as much as more mid-high priced commodity items. Low priced commodity items will flourish.

A good example of this kind of consumer business is the car industry. Not luxurious models (rich folks will still buy expensive luxury cars) but the more modest car lines will suffer enormously because they often finance the deals that people can no longer afford to pay.

We see this already with GM & Chrysler bailout by the US government. if it weren’t for the massive economic threat of losing these industries to bankruptcy (and the associated job losses/knock on effects) these giants would probably go out of business. They will still suffer huge downturns in 2009 due to reduced demand.

Businesses with no differentiation and focus on cutting costs rather than increasing profits will have a tough time.

Becoming a strong operation

In order to become a strong company you need to do exactly the opposite of what I just called weak.

• Differentiate your brand (become needed)
• Measure your advertising and overall business performance and act when you need to
• Become critical about your value to the customer and get creative with how you market that value
• Focus on more profit - where your spending produces positive ROI spend more. Where it produces less ROI reduce spend.
• Spend better not less

The Analytics industry sits in the middle of all this. Einstein once said;

At the heart of every difficulty is an opportunity

As an industry my thought is that it is going to get tough, as difficult a ride as it has in previous recessions. However what is interesting is that we’ve never been this mature before, we now have the power to really help businesses become stronger operations by cutting costs, improving marketing ROI and pinpointing opportunities.

I predict that real growth systems will evolve in 2009, creative methods to help companies grow, using analytics, using better online and offline marketing by defining what the value is to the customer and speaking to them in their language. Am I an optimist or what?

The case was really interesting because the the client, a mobile telecommunications operator based in Indonesia wanted to create buzz about their brand new service offering and kick off the usage of MMS services. The more MMS services used means more revenue for the company as every MMS message is charged to the consumers phone bill.

A competition was launched where consumers would have to use their phone to go to a mobile website, complete a number of tasks (including sending a number of MMS messages). Points were awarded on the mobile website for every task completed, the point being to try to amass more points than anyone else. The competitions top 50 would win a lot of free phone credit while the top 10 won cool new phones and the top 3 won top of the range motorcycles.

Our challenge was to measure the success of this campaign. We ran a KPI workshop and defined the KPIs. The main business success measured would be MMS campaign value and then MMS usage lift in the active users.

In the KPI workshop we broke the campaign up into a REAN (Reach, engage, activate and nurture) map.

If you click on the image you’ll see the different elements in each dimension.

This allowed us to plan how to measure the campaign successfully. The idea was to channel people from online campaigns and offline events to a mobile website or to a normal website where they could find out more information about the competition. The website listed in the engage part of the model was a campaign website designed to persuade people to go to the mobile website and take part in the competition using their phone.

We knew it would be extremely difficult to measure the success of the campaign without identifying the competition participants. Therefore when we finally got the participants to the mobile website they would have to sign up with their phone number to receive instructions on how to proceed (via text message - tracked by the Telco). In this way we could also track exactly which tasks (conversions in the activation part of the REAN model) each participant had completed and could award the points accordingly.

There were also challenges with some of the offline event locations. How did we know for instance that people attending the Telco’s live event launch had signed up from the event after hearing about the competition there or had they signed up after seeing it online? We solved this by giving each event screen a different ‘activation code’ meaning we knew which screen and which location the mobile site sign up came from.

The tasks and users were logged in the mobile websites content management system. With this campaign therefore we had a number of different data sources to measure the effect of the campaign.

• We had SMS sent/clicked from the telco (so we had a success ratio immediately).
• We had Mobile CMS log file data
• We had event screen data (logged in the mobile CMS)
• We had website data (Google Analytics)
• We had MMS traffic logs (from the telco) going back one year and showing traffic peaks during the campaign period

The results were a considerable lift in the number of MMS messages sent, many active participants and tens of thousands of tasks were completed (coming form the CMS logs). The campaign also contributed to an overall lift in MMS usage (seen in the Telco’s bandwidth traffic logs), we had a vast amount of website visitors which was a goal for the branding and a good sign up ratio. The client was very happy about the results and have since run similar campaigns.

We learned a lot from these campaigns and were able to pass the learning from this campaign to the next one for the client. One finding for instance was making the SMS sign up process very simple as people abandoned in large amounts when they had to react to more than one SMS message to activate their participation in the competition. We also saw that by not teasing the Telco’s website quickly enough we lost a lot of free traffic we could’ve taken advantage of.

Because this campaign was a mobile campaign and the measurement was multi-channel we had a number of different challenges. However what was key to successfully measuring the campaign was a KPI workshop, effective use of the REAN mapping and good coordination of the event both on the ground and online.

Comments? Similar stories?

On Thu, Sep 25, 2008 at 5:56 AM, Jackson, Steve wrote:

Hi google,

Why on earth are you discussing basically an interface change which does help, looks nice but is basically cosmetic when in actual fact you have just revolutionized the entire platform by adding powerful, customizable, behaviorally based segmentation?? GOOGLE ANALYTICS WITH SEGMENTATION! You just moved into the big league!

I nearly missed it! Advanced segmentation (beta) in a tiny little link. You just turned GA into SiteCatalyst with Data wharehouse! You just made my day! When does this feature go live? because this is what Google Analytics has been in desperate need of for the last 2 years.

It’s a HUGE move on your part.

Could you tell I was excited? My point about SiteCatalyst is of course over the top, Omniture have a widely different value proposition and you get what you pay for, but in terms of pure features GA and IndexTools are there or therabouts.

I hinted in my last post about Yahoo! that Google would respond. Well they did. I was going to write a long post about how it worked and what you should do, but Avinash already beat me to it. I’d fully recommend reading his excellent article on the matter and once again I have to tip my hat to the folks at Google.

When Dennis Mortensen first told me that Yahoo! were buying IndexTools and rolling it out for free I was delighted. Now that it’s official I’d like to congratulate all the folks involved at IndexTools again for producing a damn fine tool and Yahoo! for raising the value threshold of free Web Analytics.

I predict that soon we’ll see even more advances in the free offerings. Google will respond, this I know. However the real pressure is now on Omniture and other paid offerings to rise to the challenge and make sure that their pricing is justified by the value they offer their clients.

Cool! Kudos to Yahoo!

His answer was that he would like to have one tag that all web analytics vendors used. This is something I have given thought to before so I continued the discussion in the pub (inevitably). The debate came up with all the objections I expected which I’ll come to later in this post. It also came up with all the benefits that could result from a universal tag.

I spoke to people from Omniture and Google, people from agencies, people from consultancies, folks from content management system companies and coders (the great thing about eMetrics is it brings these folks together) and not one of them had any problem with the concept. Everyone thinks it’s a good idea.

Indeed I asked a panel at eMetrics and Alan Boydell of Keyade said they had already developed a universal tag for 1000 websites that had multiple analytics platforms running. The only way he said that they could get data for all of them was to develop a tag that sent data to one repository (as well as the individual vendor data centers).

Let’s be clear what we’re talking about and what the benefits might be. We’re talking about a very light middle layer tag that could be applied to all websites that assigns a cookie to the visitor, tracks visits and records page views. It would also have a link to a Javascript file that then routes that information to whatever web analytics vendor was used by the visitor.

The Benefits

• One tag means that basic level data collection will cut implementation costs for any enterprise.
• One tag to rule them all means standard implementation policies for content management systems, e-commerce systems, online booking systems that currently struggle with web analytics implementations.
• One tag opens up possibilities for benchmarking across industry verticals (and new opportunities for companies that want to collect universal data from opted in companies).
• One tag means the paid vendors have a much smoother adoption possibilities. If for instance a company has Google Analytics implemented one downside to changing vendor is changing tags. If a universal tag was used a simple switch could turn on the basic Omniture implementation.

The Objections

The vendors would be the people that I would expect have the most to object about.

• Patents have been registered about tagging methodologies. True. I understand the business value of the patents and no-one is asking vendors to give way their tagging technology, merely to adjust current technology to work with a universal tag.
• A universal tag would not solve anything in terms of advanced implementation and deriving business value via KPIs and such like. Again True. Advanced implementation of something like Omniture would still remain a technical challenge. But a universal tag would mean that 3rd party vendors would have more understanding of what is required and a basic tag would mean basic data collection is already being done.

So what say you?
One tag to rule them all, one tag to bind them, one tag to collect it all and from the darkness find insight?

Is it just a dream? Or should the Web Analytics Association take an action point and develop the technology? Would vendors consider using it if they did?

On Friday 19th September I will be a keynote speaker at the Design For Conversion conference in Amsterdam. Ton Wesseling A dutch Internet Marketing Consultant and Managing Partner of Orange Valley interviewed me for the dutch marketing optimization blog WebAnalisten. I thought you might be interested in my answers so published them here also; 

1. What are the lessons to be taken for the online field from old fashion 1 on 1 (sales) talks?

Understand that your website is a one on one sales talk. You’re doing the talking and a person is doing the listening (reading). Don’t therefore talk to an audience, talk to the individual. The sales folks dealing with objections on a daily basis can help your online sales process immensely. Knowing what the benefits of the product or service is, relating that in the copy and content, and giving your visitor an answer to their objections increases the chances of them making a decision to buy or order from you.

2. With all the optimizing and persuasive marketing techniques: wouldn’t the website visitor become blind for these techniques?

Optimization is simply a tactic. If your content is bad you will fail no matter how well you “optimize”. In reality why would a visitor who is looking for a solution to a problem leave when you solve that problem? Why wouldn’t they get in touch or buy your product if it solves their need? The fact is that no-one cares whether you’re optimizing or not, they just want to have their problem solved. If you do that you win. If not you lose.

3. What will make you more money? Excellent profile based offering to returning visitors or excellent offers based on session behavior?

It depends on your service or offer. I am tempted to say the profile based offering to returning visitors but then again if you’re selling a common commodity you win or lose based on the first impression. Generally profiling and segmenting visitors based on behavior gives you more insight which when applied should make you more money. 

4. Mobile internet seems to be a different ballgame: What will the knowledge of physical location mean for marketing optimization?

This is an interesting field for me working with Telecom companies as much as I do. It’s difficult currently to measure mobile marketing as it stands now and new methods will have to be introduced to plan campaigns. All marketing can be measured if the online channel is used as a proxy. By utilizing mobile vouchers for instance we could match offline actions with online sign-ups. But imagine the power of knowing that a person interested in say a certain type of music was near a store that sold it. If you’re that store owner an advert sent via SMS then and there would be a very powerful offer. What we need to do is introduce good measurement planning at the pre-marketing phase. Measurement allows us to then optimize the experience.

5. At what point does persuasion moves over the edge and becomes influencing? (and is that a bad thing?)

I don’t think persuasion or influencing are bad things, people are grown ups and make decisions based on the information they have. When it moves over the edge and becomes bad is when people con you. When messages are false, when testimonials are made up, when marketing is used to rip someone off is when it becomes bad. This however is a fact of life and why risk reversal is so important in marketing.

6. What is in your opinion the website that is a perfect example of excellent design for conversion?

None of them are perfect but here is a benchmark list of e-commerce site conversion rates.
http://www.grokdotcom.com/2008/07/24/top-10-online-retailers-by-conversion-rate-june-2008/

7. It’s all about customer LifeTimeValue. How could you really optimize your website when you can’t tell what the LifeTimeValue will be?

I have saved clients millions without RF models or any customer information by focusing on reach, engagement and activation. To give an example by setting up a measurement which looked at a segment of folks who stayed on a website for more than 1 minute and viewed more than 3 pages we were able to optimize search engine marketing campaigns to drive engaged visitors instead of just visitors. This saved marketing spend itself but more importantly we saw another campaign (from banners) that engaged this segment very poorly meaning that bad spend (in this case a big campaign costing $2 Million) was diverted to more effective reach sources.

Once you have enough data and with enough planning you can also learn the customer lifetime value. This is a step up from web analytics to customer analytics and requires data mining but it is possible.

8. How could you use all the conversational online information about your product or niche to improve your users conversion?

It’s a good source of voice of customer (VOC) data that you can use to build personas, product or service USP’s and also can be used to give insights as to how you stand in the marketplace alongside your competition.

9. Where does delivering a good user (web) experience stops and will you have to focus on converting the visitor?

In my opinion delivering a good web experience and converting the visitor go hand in hand. However what I call a good experience is excellently answering the needs of your potential customer. Website usability is important but it’s less important than answering your customers’ questions. I’m sure you have bought something online from a website where it was really difficult to get to the end of the purchase process. This would probably have put off a lot of people that were casual about the product or in a hurry and got frustrated which is why we optimize, but in the end if you really want it you will jump through all the hoops to get it. In order to get your customers to really want it you have to have somehow addressed all their needs. Bad usability is rarely a show stopper for the customer who is 100% convinced you’ve got the best deal for them.
�
10. If you find out that your visitors will have different optimal (graphical) conversion designs during weekdays or hours. Should you present these designs, or will the different graphical designs be too negative for the total user experience?

It would again depend on the product or service. On valentines’ day I know of a flower delivery company that monitors their website homepage in realtime and adjust the images accordingly because it’s the biggest day of the year for them. They monitored what people were searching for in the Google Keyword Tool so that they could react accordingly and then they changed their offer every time the conversion dropped below a certain level. It worked very well, so yes I would advocate this kind of testing for sure. Planning is key, be careful to balance what the customer needs are with why you’re making the change. I wouldn’t change the site layout for instance or the general navigation, but individual imagery and offers shouldn’t be a problem for most people.

His job is to help Omniture build their business in this part of the world. This is positive news for Omniture and for Trainers house as Omnitures biggest partner in Finland.

Kalle will step down from his role as the Web analytics associations country manager of Finland after we have found a suitable replacement. We intend to build a board of directors for the Finnish WAA and Kalle will continue to run the Finnish operations until the new board we install votes in a new country manager (to act as a Finnish Chairman and report to the International committee)

I’ll look forward to continue working with Kalle in the future.

That at least is a step forward. At least some authority in the UK has been officially assigned the responsibility to take the complaints on board. When I made this complaint in June they were playing catch with the UK Home Office as to whose responsibility it was. The EU comission ends the letter with an encouraging;

The Commission is currently in contact with the UK authorities to clarify in particular the actions of the competent national authorities with regard to the users’ complaints about trials of Phorm technology by BT in 2006 and 2007, as well as the position of the UK authorities regarding the planned future deployment of the Phorm technology, in particular the way in which it is planned to obtain the users’ consent. The commission will continue to follow this case and take appropriate action, should the need arise, to ensure that the relevant EU law is effectively implemented by the UK authorities on this matter. (bold emphasis mine)

I congratulate the EU on recognizing the problem and picking up the phone as well as threatening more if something is not done. I have also drafted another letter which I have issued to the Advocacy committee of the Web Analytics Association which is currently under review.

This is a recent excerpt I posted on WebAnalyticsDemystifieds’ Future of blog talking about the same thing (engagement);

It goes back to a point I first read about ages ago on Eric’s blog (late 2006) and then when Eric published his engagement formula became the legendary “engagement” debate I’m sure we all remember, on Occam’s Razor, Jims site, my site and a bunch of others. It got quite heated at times as it should. Passions were ignited and people were drawing lines in the sand. At the time I took a step back and looked at what we all were saying and came to the conclusion we were largely debating semantics though we all agreed on some things.

What we pretty much all agree on is that the more ways we have of identifying ways to get more customers and take actions on our metrics the better. Note the “take action” part. I think most of us agree if you can’t act on the KPI then why use it.

We all agree that we have to use quantitative (clickstream), qualitative (voice of customer, attitudinal) and competitive (comparison) data to drive the best insights. Learning to combine these data sources is the way the industry will move.

In the “legendary debate” Avinash argued the same point as Matt makes and I’m not disagreeing with either perspective here. The whole point of everything is conversion eventually.

The problem was one guy was discussing engagement in the context of RF models. One was discussing engagement in terms of bounce rate, one was discussing engagement in terms of scoring actions and I was defining engagement segments. Then everyone else was throwing their 2 cents into the mix till you had a whole big bag of ideas labeled as engagement.

I have used Omnitures (and other) tools successfully to save clients millions of euros using nothing but engagement segments as I define them. No conversion was made initially but the likelihood to convert later (measured via RF) was much higher and by using engagement segments to optimize keyword spend and banner placements we saved clients a lot of money.

As a consultant using (Omniture and other) tools it’s our responsibility to make or save our clients money. I often find that in order to do that I have to do a lot more than simply measure conversion funnels because all that does is show low hanging fruit which will only give you so much (and only works for some sites).

The puzzle I am being asked to solve more and more is this.

How come multi-channel campaigns for the past 12 months have shown only average or unremarkable conversion results and yet our profits are still rising exceptionally? Which part of our ad money are we wasting? Something is working but which is it?

In the above situation conversion rate alone is worthless. Much like engagement index alone, or bounce rate alone or page views alone.

The point I think folks might be missing is that it’s not about “one engagement metric to measure all” it’s about taking various metrics in context to each other and knowing how to interpret and act on those metrics.

Till someone replies!