Getting The Most From Paid Search In A Difficult Economy, Part 2: Valuation Models

How much revenue does a new customer drive to your business? How long, on average, does a customer last, and how much do they spend as a customer? Do customers acquired through one marketing channel generate more revenue than those from another channel? The answers to these questions collectively make up the foundation of any direct marketing effort, and are especially important in paid search, where every click costs you money. These days you need to know the answers or you may quickly find yourself without a marketing budget.

In my last Industrial Strength column, Getting The Most From Paid Search In A Difficult Economy, I wrote about paid search management tactics and the different business goals they can support. We looked at a few ways to think about profitability, depending on whether you are in growth mode or you are looking to be ultra-profitable on your last dollar spent. My point was that paid search advertising is unique in its flexibility, its ability to “turn the dial” up or down depending on what your business goals are at any given time. Today I’m going to take a look at another dimension that greatly affects advertisers’ paid search tactics: customer valuation. When I speak at industry conferences, I often talk about—and get lots of questions about—valuation. I preach about the fact that this is where all good direct marketing starts: understanding the value of your customer or user. I also wrote about it in a previous column on SEO valuation, but it holds true for SEM as well. Valuation is important, not only because it is the foundation for any sound marketing campaign, but also because, like profitability, it is just another dial you can tweak to better support your business goals, even as the business climate shifts.

At Yahoo! we have a concept of “lifetime value” (LTV). We get LTV by modeling the net present value of the revenue stream a user creates on our site. (Note: Someone asked me about how to generate an LTV recently and I was delighted to find out that Excel has a net present value formula that makes it really easy – it’s “NPV” – check it out!) At Yahoo! we have different LTVs for different products and services we promote, as well as different LTVs for different marketing channels we use to promote these products and services. For example, for Yahoo! Personals subscriptions, we might use an LTV of $100 for SEM, but for our Affiliate channel it could be $89, it just depends on how the subscribers’ revenue streams differ (how long they stay subscribers, which package they buy, etc). This means that, all other things being equal, in this example we would have a slightly different target cost per acquisition (CPA) for SEM than Affiliates when promoting Yahoo! Personals subscriptions.

I like the notion of LTV because it helps level the playing field for all marketers. It gives us a common language to speak and a universal set of metrics by which we can evaluate our success. We need this type of tool because we have many different businesses with very divergent business models. In our example above, Yahoo! Personals is trying to generate subscriptions. Yahoo! Small Business, on the other hand, sells domains.  Yahoo! Autos captures leads for automobile dealers, Yahoo! Shopping supplies links to merchants with great deals on products, and Yahoo! News serves ads to its readers. Tools like LTV help us measure the value of marketing programs across all these properties and across all marketing channels.

Before we dive into other valuation models, I want to point out that valuation is one of those things that we do at Yahoo! because our scale and complexity demands it. At the same time, it’s something all marketers should do, even if it means sacrificing sophistication for simplicity. As noted above, there’s even an excel formula to help you get there, so there’s really no excuse not to take a swing at it. And, as I always say, if you don’t know what the revenue model looks look like, take your best guess. Just remember to go back and challenge your assumptions later.

LTV works great for marketers and businesses that want to be aggressive with their marketing tactics. It allows you to leverage the full value your customer brings to you in your marketing efforts. Much like we talked about startups using a portfolio approach to managing paid search, a company with a similar risk profile would want to market to the full value of the revenue stream their users create.

But what if your company is not in an aggressive growth mode? What if, due to macroeconomic conditions, you need to be more conservative? What if you need to focus on being profitable in a shorter timeframe than the full lifetime of a customer? In my last column I mentioned that a conservative way to manage paid search was to manage to incremental ROI rather than average ROI. Similarly, a conservative strategy here would lend itself to a valuation model more conservative than LTV.

At Yahoo! we have a number of revenue metrics that are based off the same valuation models that bring us LTV. LTV, for example, has a finite horizon of several years. From there, we can dial back the horizon to suit our business needs. One metric that we find very useful these days is called rolling twelve-month (RTM) value. RTM helps us understand what a customer is worth to us in the next year, regardless of when the user is acquired. RTM works well because, while it is a more conservative valuation than LTV, it takes into account a full year’s cycle, so seasonality is not an issue. From RTM, the next conservative step is in-year value. In-year looks at the revenue stream from customer acquisition to the end of the business year (so it changes dramatically based on when in the calendar year you acquire the customer). This helps advertisers focus on how much value is driven within the fiscal year. Although somewhat shortsighted, it makes sense for public companies to look at this valuation, because such companies are accountable to Wall Street for earnings goals.

For an even more conservative view we can look at rolling 3-month and in-quarter valuations. These short time horizons don’t allow much flexibility in direct marketing goals, but they do give a very finite view of how much value a customer will bring in the immediate future. Direct marketers may not like this model much, but finance and business folks like to hold marketing programs up against these valuations to make sure they’re doing the right thing for the company, no matter what valuation they’re using as an ultimate measure of their success.

Finally, in some businesses, the very nature of the revenue stream is very short-lived. Examples of this are where users come into a site, consume content, and click out on a link to go to another site. In cases like these, we will sometimes throw the above models to the wind, and even look at direct in-session revenue. While this is clearly the most conservative valuation of a customer, it brings with it several key advantages. First, it’s easy to model if you can track it on a very basic level. Second, as the economy fluctuates and businesses are wondering whether to use LTV, RTM, in-year or in-quarter valuation, the direct revenue model is already so conservative that advertisers using this method are relatively unaffected by the macroeconomic climate.

So, just as profitability metrics provides one set of controls that affect how we manage SEM campaigns, valuation methods provide a whole other set of dials advertisers can tweak to excel in a rapidly changing economic climate. As the business climate continues to shift, you will want to use all of your options to create the most appropriate marketing programs relative to your business goals, whatever they may be.

David Roth is Director of Search Engine Marketing for Yahoo!, Inc.. The Industrial Strength column appears Mondays at Search Engine Land.