De-averaging your marketing investments and paying close attention to how your incremental marketing investments pay off is a post I've been intending to write for a while now.

The good folks at Rimm-Kaufman group, however, saved me the trouble with yesterday's excellent PPC Averages can Hide Incremental Nightmares.  Please take the time to read it.

As I was reading it, I started to think about applying LTV to the methodology.  When you think about making decisions based on LTV, the story becomes even more clear.  All will be revealed (with a graph!) after the break.

First, I roughly recreated the RKG dataset from visual examination of the graphs.

Then, I developed some hypotheses about how the customers acquired in the original effort would perform.  They included:

1. Incremental Lifetime Sales. The top bucket of customers (acquired with the first $10,000 in ad spend) would generate an additional 80% more sales.  The bottom bucket would generate just 5% more sales.
2. Incremental Lifetime Marketing Investment. The incremental advertising cost to generate the incremental sales (you have to spend money to make money, regardless of what your CFO may want to believe!) ranged from a 7% A/S ratio for the top bucket to 1,000% for the last bucket.

I suspect I've been too conservative in my assumptions of incremental sales for the top bucket and the rest of the A/S assumptions are just estimates for discussion purposes.  I made no additional assumptions for improvements in gross margins and so on.

So what happens when I applied my assumptions? Here's the graph:

The light yellow and light blue lines are essentially the graphs on the original RKG blog.  The orange/dark blue lines show what happens when I try to add in LTV.  What happens?

1. Peak profit (marketing income) shifts leftward. That means that instead of hitting max profit at around $60K of original acquisition ad spend, this model shows it peaking at around $40K of original acquisition ad spend.
2. The profit slope changes. It becomes more steep in the first four buckets of marketing investment and then the rate of change also falls off more rapidly after the peak.

Of note is that I haven't considered overall revenue and other factors (economies of scale) that may impact marketing investment decisions.  You may, for example, want to be slightly less efficient on the margins for revenue volume purposes or to maintain volumes if you have costs that operate as step functions.

What does this mean as you look at incremental marketing investments?

1. It's even more important than Rimm-Kaufman indicates when you look at investment on an LTV basis.  The mistakes on the margin are not as gradually dilutive to your profit, when you consider LTV.
2. It's important to get up the experience curve. Given that the slope of profit rises pretty quickly, you're leaving money on the table if you don't (or can't) bucket the marketing investments and quickly march up that curve--by investing more in marketing, smartly.

How do you analyze the incremental performance of your PPC campaigns and other marketing investments?  What challenges do you face when trying to build models to maximize your marketing performance while delivering against volume or revenue constraints?

Update, August 6th--Graph above was updated due to some missing data.  All comments above still hold.  The original graph uploaded this morning is available here for review.

Bookmark to:

At one point during my college days I wanted to be an A&R guy for a record label.  My reactions to smoke-filled clubs and early-to-bed habits caused me to rethink that career option.

But music, and the marketing of it, has remained a lifelong interest.

Last night, my wife and I saw Duran Duran at Merriweather Post Pavilion in Columbia, MD.  This was her 11th or 12th time and my 10th time to see the band.

When you go to a Duran Duran show, you know you're going to see a great performance, an enthusiastic crowd and hit after hit.

What I didn't expect was a textbook example of creating and maximizing a marketing channel, and an example of how big a part luck plays in everything we do as marketers.

Creating a marketing channel.

Opening for Duran Duran was Your Vegas, a band from Leeds, England.  We sat down early to catch a few minutes of their show and thought they were pretty good.

At the end of their set, they thanked everybody for listening and said they'd be up at the top of the hill signing autographs and giving out free posters.  I listened to their call to action and headed up the hill to see the response, and purchase a CD.

I was pleased to see no aloof rock stars there.

That call from the stage created a long line in front of the entire band, where they took their time chatting with the fans, signing the free posters and CD's and taking photographs with the fans.  The band was gracious, appreciative, and seemed to be genuinely enjoying the interaction.  Further, their new CD was moving briskly from the stand next to them.

Will Your Vegas be the next Killers?  Luck will play a big role, but the band is doing everything right by:

1. Creating a face-to-face marketing channel, starting with their call to action from the stage
2. Engaging in CRM by getting face-to-face with 50-100 fans every night.
3. Making a small investment in posters and pen ink to support the campaign.

Luck and music.

Anna Ross, Simon Willescroft and Dom Brown joined the band as backup vocals, sax and guitar, respectively.  All are highly talented and excellent performers.  Ms. Ross has been touring with the band since at least 2005 as well.

Any of these folks could be headlining a show with their own band.  Why not, then, last night?

In a word, luck.

• Answering the right ad in a paper looking for a guitar player.
• Sitting next to the right person in grade school.
• Being in front of that guy in an audition.

Anna, Simon and Dom are surely doing fine touring with Duran Duran and have done everything right to prepare for success.  One or two breaks and they might be headlining.

The point of this?  The impact of luck on your marketing efforts can't be entirely eliminated, but can be minimized by proper planning and preparation.  Also, don't get down when things break the wrong way, as will invariably happen.  Go right back to planning for the next effort.

Summary and key takeaways.

1. Look for unique opportunities to talk with your target audience.  Don't necessarily think "sales channel", but instead find unique opportunities to talk with those who are interested in your products and services.
2. Prepare for the impact of luck.  Preparation won't eliminate the impact of luck on your campaign outcomes, but it will position you to capture the upside when things are working well.
3. Wear earplugs. Tinnitus (of which I have a mild case) is no fun and only gets worse over time.

Bookmark to:

How do you make money?

This is one of the first questions I ask all my clients. The answer usually comes back including some aspect of "buy low, sell high" and other margin-related facts. Regardless of the complexity or depth of the answer, one word is always included.

The magic word is "people."

So why isn't all marketing done on customer lifetime value (LTV)?  What are the five things you need to consider when developing LTV-based models that allow you to build CPA (cost per acquisition) based marketing plans?

You typically don't make money on the one-shot sale.  Your automobile dealership didn't make money strictly on the sale of your car, unless you struck an extraordinarily poor deal.  It was the financing, insurance, trade-in and accessories that made the transaction palatable to them.

The real money is in the service and the 2nd and 3rd cars that you might buy in the future.  (Along with the ancillary financing, insurance, trade-in and accessories associated with those sales.) Acquiring you as a customer is the key to long-term profits for the dealership.

Why doesn't everybody consider an LTV-based marketing model?  

I was speaking to a potential client last week and asked them "how much did it cost to acquire new customer?" and "what is the lifetime value of your customers?"

The fact that they didn't know the LTV of their customers wasn't a surprise.  It was a new business, after all, and they didn't have any history.  What was surprising was that they didn't believe customer LTV and having a solid CPA was the way to look at the business.  Marketing was not looked at as a revenue-generating component of the business. Rather, it was seen as an expense and nothing more than a percentage of sales.

I thought this was interesting. Is it true that there are some businesses where looking at customer acquisition costs and calculating customer LTV doesn't make any sense?

I thought about it all weekend and came to the conclusion that--in the B-to-C space--the answer was no.

If you generate profits by selling to customers, then you need to calculate and understand customer LTV and use that to understand your allowable CPA.

Estimating (or calculating exactly) customer LTV fundamentally only includes about five factors.  They are:

1. Understanding your conversion funnel.  You have to have some idea of the Suspect to Prospect to Lead to Buyer flow. You have to know that conversion rate, because your marketing activities will bring in people in different parts of the funnel. Sometimes, your marketing effort will get you no more than a suspect. Other times, such as in SEM, you'll get somebody who's ready to buy right then. Knowing and understanding the flow of suspect through buyer will help you allocate your marketing dollars more effectively across all your channels.
2. Initial unit of sale and margin on the unit.  You need to know what the initial unit of sale is and the level of profit (before marketing expense) is included in that unit of sale.
3. Opportunity for repeat purchase.  Is your product one that needs to be replenished periodically?  How frequently?  How many of the initial buyers will replenish from you, as opposed to your competitors?
4. Ancillary purchases.  What other products and services do you sell?  What is the attach rate for each of those to your new customers and installed base?  In our case of the automobile dealership, what percentage of the new car buyers will choose to come back to the dealership for routine maintenance?  (And what kind of marketing activities will you have to undertake to encourage them to do so?)
5. LTV window. This is probably one of the most important things to consider.  In the case of our automobile dealership, the lifetime value window for a car buyer in her 30's might be 5 or 10 years.  In the case of a digitally-delivered product--say, music downloads--I think you'd be taking on a lot of risk to consider a window longer than 18 to 24 months.

The rest is just building an Excel model to determine how much you could spend and then developing a bottom-up plan to exhaust all possible marketing channels that can deliver against that allowable cost per acquisition.

Yes, sometimes you'll have to make some assumptions and sometimes you'll not have all the data you'd like.  But if you know the ballpark you're playing in, you'll know which marketing channels might make sense and make your ROMI goals.  You'll also know where to take calculated risks to stretch your marketing budget and grow more quickly.

Summary and takeaways

1. Every business should know the value of a customer over some reasonable time frame, because profits come from selling products and services to customers.
2. It's relatively easy to calculate customer LTV.  Even if the data isn't entirely clean, it's still pretty easy to determine at least the rough value of a new customer.
3. Build your acquisition marketing plans using a known allowable CPA.  Once you know how much a customer is worth, it's easy to figure out how much you'd invest to get them to buy from you.

So what do you think?  Can you really calculate customer LTV for businesses as diverse as magazine subscriptions, convenience store customers and shoe store customers?

Bookmark to:

A client recently asked me "who can tell me if my website is working well for me?"  My immediate response was "your customers and your browsers."  This, of course, triggered a conversation of how it was possible to talk to tens of thousands of (usually) anonymous visitors, collect their insights and then translate that to marketing improvements.

Prior to a large amount of advertising moving to the web, with the associated tracking and analytical capabilities, my response didn't make a lot of sense.  Unless you were the sole proprietor of a local general store or had massive resources to undertake a large amount of expensive primary research, it was really hard to figure out what exactly about your marketing was working for your customers and prospects.

Until the rise of web analytics, that is.

In my opinion, the best thing that's happened in the last several years for online marketers was when Google acquired Urchin software.  This event made it easy and free for marketers to analyze the results of their online marketing activities.  Sure, for large business or those with specialized needs, solutions such as Omniture might be better.  But the majority of businesses can do just fine with Google Analytics.

And now that Google Analytics was free, it became fairly simple to add another great product: Google Website Optimizer.  This product makes it easy even for a small company to undertake A/B or (provided the traffic is heavy enough) multivariate testing.

If you've found yourself in any meetings recently arguing the merits of "red vs. blue" or where to place a button on the website, ask your online marketing team if they're using Website Optimizer or any other systems, such as Optimost or Offermatica.

If not, get one of these installed now--life's too short to spend time arguing over the color of the sidebar, the placement of a button or the length of the copy on a website.  Especially when your customers will tell you exactly what makes them buy from you.  And when your internal back-and-forth will result in you making decisions that result in suboptimal ROI.

If you are using one of these optimization products and you're still having these meetings, then something else is wrong and you may not be using them correctly.

Trust me, your customers--when it comes to buying your products and services--are far smarter than you when it comes to landing page and website design.

I used to teach an internal class on direct response marketing and would show an example of three different online creatives, with the objective being to drive downloads of a software product.  When I asked a room full of 30+ people that probably contained over 400 years of marketing experience and probably 35 advanced degrees which creative delivered the best ROI, less than 25% picked the winning creative.  And even those folks couldn't tell me why they voted for that one.

Now imagine a multivariate test of a heavily-trafficked landing page, which results in several thousand "recipes."  Any chance your CEO picks the one that sells the best?  OK, maybe they hit it right once--blind dogs and acorns, after all.  But how about the nth time?  Not a chance.

So, let's say you're not using Website Optimizer or a similar product and maybe aren't too familiar with the products.  What to do now?

First, I suggest watching a quick (5 minutes or so) Google Website Optimizer overview.  Second, find an hour or so and watch this YouTube overview of the product, which goes into more detail about how to use the product for an A/B test.

Then, test something.  Run a simple A/B test of a couple of your landing pages.  Or maybe your registration pages.  It doesn't matter, as long as you get going.

Summary and takeaways

1. Don't guess--test.  If you don't empirically know why something works on your website, develop a hypothesis, design and test and know it.  Then, repeat.
2. Take advantage of free tools.  Like Google Analytics, Google Website Optimizer is free.  The only thing you pay for is your time and--by stopping the endless meetings and arguments and associated meetings--that's even free.
3. Build an organized testing plan.  You still can't test everything, unless you've got unbelievable amounts of traffic.  So definitely do use that experienced team to help prioritize the testing agenda

Bookmark to: