Monday, August 29, 2011

Insights from Eloqua's IPO Registration Statement

Summary: Eloqua's registration statement offers new and interesting details about its business. My analysis is below. Hopefully it's accurate -- I think that SEC rules prevent them from commenting if it's not.

Eloqua last week filed for an initial public offering of its stock. The registration statements accompanying these filings are like a striptease: often more interesting for what they hide than what they reveal. Eloqua’s was no exception.

Let’s start with what they showed. This included the basics: revenue, customers, profits, and cash flow.

Revenue increased roughly $10 million per year from 2006 through 2010, which is nothing to sneeze at although the higher base meant the percentage rate slumped significantly in 2009 and 2010, to about 25%. Year to date in 2011, growth is back up to nearly 40%, which would translate to $70 million for full year 2011 if maintained.

Client counts are reported only for 2008 through mid 2011. The first bit of news is that revenue per client was virtually flat during that period. I expected it to grow as Eloqua sold into larger accounts. Eloqua doesn’t offer an explanation, but my guess is heavy competition forced them to reduce prices during this period.

Operating expenses grew sharply through 2008, nearly outpacing revenue. The company tightened its belt in response to economic conditions in 2009, in particular by reducing marketing and sales costs by nearly $4 million (more on that later). It reinstated most of that budget the next year, but we can wonder how much the cutbacks slowed Eloqua’s revenue growth. The industry as a whole was almost certainly grew faster than 25% in 2009 and Raab Associates estimates it doubled in 2010.

On a brighter note, cash from operations has been positive since 2009 and looks great for the first half of 2011. The difference between positive cash flow and negative profits is one of those accounting mysteries I won’t claim to fully understand; as near as I can tell, it’s due to a combination of advance customer payments (which won’t be counted as revenue until they are earned) and non-cash expenses in paid in company stock.

So much for what Eloqua said. (Of course, these are just highlights – the full registration statement runs over 200 pages.) But what does it mean?

Ultimately, investors want to know whether Eloqua will ever show a profit. Companies in the early growth stage rarely do, so the losses to date are not necessarily cause for concern. Software-as-a-service vendors in particular love to point out that they get a recurring revenue stream from each customer, so there are future profits that are masked by first-year sales expenses. They also argue that much of their operating cost is fixed, so the incremental cost of servicing new customers is quite low. In both cases, then, present losses are a harbinger of future profit.

Let’s start with the servicing costs. Eloqua reports separate revenues for subscription and support (what I’d call servicing costs) and professional services. It also breaks out costs for those two categories. Professional services are under 10% of revenue and they jump depending on whether Eloqua performs them in-house or shares them with partners. They also run at a loss, which means they are more like a sales cost. In any case, let’s ignore them for now and focus on servicing costs.

What we see is that, instead of falling, the amount of revenue spent to support each client has actually increased as Eloqua added more clients. The good news is that it has pretty much stabilized around $45,000 per client or 20% of revenue in 2009 and 2010. The reason seems to be a substantial upgrading of Eloqua’s infrastructure in recent years, perhaps combined with competitive pressures that have prevented it from raising prices as it serves larger clients. A reasonable conclusion, at least for the short term, is that Eloqua’s operating margins won’t get much better as it grows.


Ah, but what about sales costs? After all, the company pays those just once per client, while it earns its $45,000 per client each year. Once it earns back the sales cost, that future contribution is gravy (or, at least, contribution to other fixed costs).

And what is each sale costing? The registration statement shows marketing and sales costs each year and year-end client counts. So we can easily calculate the number of clients added each year and, from that, the sales cost per new client. What we find is a rather frightening $120,000 per client in 2010, and about the same so far in 2011. At $45,000 contribution per year, that would take nearly 3 years to recoup. Ouch.

But that’s assuming all sales and marketing costs are spent on new clients. Some of them probably go to service existing clients – although that actually makes things worse, since those costs would recur each year and reduce the operating contribution. What makes things better, sort of, is recognizing that Eloqua actually loses some existing clients each year.

This is where what Eloqua didn’t reveal is intriguing. I couldn’t find retention rate anywhere in the registration statement. This is a critical variable in software-as-a-service economics and is regularly reported by public companies. I’m not aware of a rule against Eloqua reporting it, although I’ve never researched that point. (If anyone out there knows either way, please tell me.)

In the absence of real information, I’ll just blindly speculate. Let’s say Eloqua loses 20% of its customers each year – a figure that translates to better than 98% per month retention, which would be considered pretty good. We use that value in two ways: to calculate the number of clients Eloqua must sell to replace its losses, and to estimate the lifetime value of a client.

Adding in the replacement clients, the cost per new client comes down to about $70,000, for a payback of about 1.5 years at $45,000 contribution. That’s pretty respectable.

Attrition also reduces the lifetime value: at 20% per year, the five year value comes to about 3.4 years (1.0 + .80 + .64 + .51 + .41), or $151,000. Subtracting the sales cost, you still have a nice $81,000 contribution to other fixed costs over the five year period. Those other fixed costs (R&D and G&A) have been rising much more slowly – around 15% per year, to where they are now around 36% of revenue.


If we put all those figures together, assume a 30% growth rate, and ignore professional services, it looks like Eloqua should just about break even in 2011, make a reasonable profit in 2012, and make a substantial amount in 2013.


But can we really ignore professional services? As I mentioned earlier, it runs at a loss but has varied greatly over time: from recovering 91% in 2007 to just 51% in 2010. The closest relationship I can find is between professional services revenue and marketing and sales cost: both were cut in 2009 and have rebounded since.



A somewhat optimistic estimate of professional services losses might be the actual rate for 2011: revenue at 20% of sales cost and recovering 60% of expenses. At that level, my estimates push out profitability to 2013. On the other hand, Eloqua has considerable control over pricing of its services. Should market conditions improve, they might be able to raise prices again to something close to breakeven. So perhaps I’m being pessimistic.


Back to that earlier question: What does it all mean? Certainly that Eloqua can be profitable even at current levels of pricing, operating costs, and sales costs. Not every firm in the business can say that. Equally important, I suspect improvements are possible in each of those areas – and would fall straight to the bottom line. So, current losses notwithstanding, Eloqua's future looks pretty bright.

Monday, August 22, 2011

Affiliate Summit East: How to Stand Out in a Crowded Market

I spent several hours this morning at the Affiliate Summit East in New York. This is a corner of the direct marketing industry I haven't examined in depth although it overlaps with lead generation, online advertising, and performance measurement segments I know well. The exhibit hall was enticingly crammed with unfamiliar vendors, although I saw enough to identify a few general categories: offer aggregators who bring together products for affiliates to sell; network providers who assemble audiences for affiliates to sell to; and technology providers to support both groups.

Swimming through the aisles were the affiliate marketers themselves – people who make their living by selling other companies’ products in return for a share of the revenue. They are the main audience for the show and, thus, of the speakers. I heard two presentations: a keynote by Wil Reynolds of SEER Interactive and a session on starting an affiliate business by Nicholas Reese of Microbrand Media. Both shared the somewhat disconcerting theme that affiliate marketers shouldn’t build traffic with shady techniques – something each speaker admitted, with the fervor of a repentant sinner, he had done in their early days. Each also offered ethical alternatives for creating a sound business.

Reynolds offered nuts-and-bolts advice, including ways to gain new links, create more appealing search results, and attract social traffic. He also offered recommendations for using a slew of free or low-cost tools, including:

- export.ly, which analyzes your Twitter followers
- Rapportive, which assembles profiles on contacts
- SEOmoz, which offers a package of features including lists of people who link to you
- Row Feeder, which reports search results against Twitter and Facebook
- Speechpad, a $1 per minute voice-to-text transcription service
- Promedia Suggester, which suggests search advertising keywords and topics

Reese worked at a more strategic level, offering advice on how to identify a promising affiliate business opportunity and then how to develop it. He made an intriguing distinction between problems solved with the customer's money (good opportunities) and problems solved with the customer’s time (not so good, because it’s harder to generate transaction that will yield a commission). He also suggested careful consideration of customers' knowledge (enough to recognize the problem exists, but not so much that they already know how to solve it). And he also listed churn rate as a key factor – he didn’t elaborate on that one, but I assume his point was you want customers who’ll stick around enough to generate repeat business.

For building a business, Reese argued that email is still king because it’s much harder to ignore than social media. He also urged marketers to build unique strategic relationships with companies that don’t have standard affiliate programs, since those relationships have higher profit potential than a program that’s already open to others.

That last point is important: with so many vendors providing packages of products to sell and of audiences to reach, it’s easier to start an affiliate business but harder to offer anything unique. Skillful content generation and traffic building may still be enough to build a profitable business, but I suspect that large-scale success requires something everyone else can't buy off the shelf. This might be an actual product, marketing technology, or business method. Affiliate marketers or suppliers who can create these will probably be the big winners in the long run.

Tuesday, August 16, 2011

Can CRM Add-Ons Replace Marketing Automation?

I’ve long believed that B2B marketing automation is just a passing phase: that, ultimately, B2B marketing automation systems will be absorbed into CRM systems instead of operating independently. It’s a view I discuss sparingly in public, since so many of my friends in the marketing automation industry have a vested interest to the contrary. But there are also a few vendors who have bet in favor of merged systems, so it wouldn’t be fair to ignore their view entirely.

Vendors have made the bet by building marketing automation add-ons to a CRM system instead of building a stand-alone marketing automation product.* I wrote in February about ClickDimensions, which adds advanced email campaigns and Web tracking to Microsoft Dynamics CRM. Of course, the jackpot here is Salesforce.com. At least two vendors have tried to hit it: Predictive Response and BizConnector, whose product is Lead Follow-Up.

Of the two, BizConnector’s Lead Follow-Up is more tightly focused, offering primarily a rules engine that allows lead nurturing and other workflows. It also provides real-time alerts and landing pages. Predictive Response has a broader range, with more advanced email features, including split tests, content templates, and branching campaigns, as well as Web visitor tracking, lead scoring, dashboards, and some data cleansing on form entries. Neither matches the scope of even a mid-tier marketing automation product, but many of the “missing” features would be available in other Salesforce add-ons. So the real question is whether the core campaign management features are adequate substitutes for a marketing automation product.

That question has no simple answer. Lead Follow-Up would probably suffice for small organizations, although it’s targeted more at individual salespeople than marketing departments. Predictive Response (which I have not examined in depth) might serve small and mid-tier marketing groups but probably not large enterprises.

Beyond these particular products lies the larger question of whether their CRM-based approach makes more sense than the separate-but-synchronized platforms offered by most marketing automation vendors. Each method has advantages:

- separate systems can present specialized marketing functions without getting in the way of sales activities.

- separate systems can use database designs optimized to process the entire database at once, instead of the one-record-at-a-time transactional processing needed for sales and service interactions.

- separate systems give each department complete control over its own system, something both groups often prefer.

On the other hand….

- unified systems avoid the need for data synchronization, which reduces complexity, eases cross-department coordination, and eliminates most risk of inconsistency.

- unified systems simplify deployment and support, especially from the perspective of the IT group (although this should be a minor factor in Software-as-a-Service systems, since most work is handled by the vendor).

- unified systems should save money, although it's not clear they do in practice.

The main technical issue here is whether marketing and CRM systems need different database structures. But even if that’s true, one vendor can provide and synchronize both structures at least as easily as separate vendors. Similarly, if separate user interfaces are really necessary, one system can offer both. So the technical differences are largely irrelevant.

Instead, the argument for separate systems really comes down the political desire of each department to have its own system, and a somewhat related expectation that a vendor focused exclusively on marketing automation will build a better system than a small division within a CRM company.

I actually accept those arguments. Vendors serving either marketing or sales will probably do a better job than vendors trying to serve both. But I don’t think this will matter in the long run. As marketing automation requirements become better understood, the CRM-based products will come closer to meeting marketers’ needs. As the functional gap narrows between the two sets of products, the convenience and cost advantages of a unified system will swing the balance in their direction. This will be reinforced by the business need for tighter coordination between sales and marketing, as well as the larger role played by corporate IT groups in selecting customer-related systems. Only marketing departments with very sophisticated needs will be able to justify buying a specialized marketing automation product.

In short, the merger between B2B marketing automation and CRM seems inevitable. In the B2B world, where sales is generally the dominant department, this means that CRM will encompass marketing automation rather than vice versa. The mechanics of the process are less predictable: CRM vendors might expand their features incrementally, acquire add-on systems, or buy a marketing automation product and integrate it. Different vendors may take different paths and move at different rates. But however they get there, I think the destination is clear.

_________________________________________________________
* For very small companies, vendors including Infusionsoft and OfficeAutoPilot have made the bet by offering their own CRM / marketing automation combination.

Monday, August 08, 2011

CRM Evolution Conference: Social CRM Takes Center Stage

I caught an all-star panel on social CRM today at the CRMEvolution conference in New York. Up on the dais were Jim Berkowitz of CRM Mastery, Esteban Kolsky of ThinkJar, Brent Leary of CRM Essentials, Ray Wang of Constellation Group, and Denis Pombriant of Beagle Research Group. You won’t find a more distinguished set of gurus, and moderator Berkowitz asked appropriately incisive questions.

Yet my final impression was that social CRM is a problem in search of a solution. Maybe it was because the speakers kept emphasizing the need to define strategy and goals for any social CRM project. Of course this is sound advice, for social CRM or anything else. But that the speakers kept repeating it suggests pretty strongly that they’ve seen lots of people deploy social CRM without that sort of planning. In addition, the goals the panel cited, collaboration and customer intimacy, struck me as pretty darn vague themselves. A couple of audience questions asking for more specific examples of the benefits didn’t yield much more substance.

The group did better when it came to tactical advice. Two suggestions that stuck with me were that companies should start with internal collaboration projects before trying to collaborate with consumers, and that companies recognize they shouldn’t be equally transparent about everything.

The panel also recognized the serious privacy issues raised by social channels, and the danger of appearing to be creepily aware of everything customers have done in different public forums. The focus was more on the dangers of displaying the information than on the dangers of gathering it in the first place. That probably mirrored the concerns of the audience.

For what it’s worth, I did pick up a pair of intriguing buzzwords: the need to display information in “context” and to pick the context based on the user’s “intent”. At least two people discussed these, so apparently they've been echoing through the gurusphere for some time without my noticing. The general idea is that context and intent can tame the flood of raw information by giving users only the data they need in formats that make sense. That's certainly not a new idea, but if some nice fresh buzzwords can get people to do a better job presenting data, I'm all for it.

Tuesday, August 02, 2011

Act-On Buys Marketbright Assets

Act-On Software announced last week that they had purchased the assets of Marketbright, a pioneering marketing automation vendor that has struggled in recent years. Marketbright’s problems have been obvious for months, so that they would vanish is not news. The interesting question is why Act-On made the purchase.

I discussed this last Thursday with Act-On CEO Raghu Raghavan. Part of his answer was because they could: having raised $10 million in June, Act-On can now consider acquisitions of less-well-endowed competitors. Beyond that, Raghavan cited several narrow benefits, including the chance of converting Marketbright customers to Act-On, the business knowledge and skills of key Marketbright employees, and some design features of the Marketbright system. He made clear that the actual Marketbright software will not be merged into Act-On because they are built with different technologies.

In other words, the Marketbright acquisition was a tactical measure to pick up some modest assets at a modest price. It is neither central to Act-On’s current strategy nor a shift towards a new strategy.

That strategy has remained remarkably consistent. It boils down to providing an easy transition between basic email systems and full-scale marketing automation. Raghavan said that Act-On has now deployed a consistent sales process to convince marketers they can benefit from replacing multiple point solutions with a single Act-On installation, even without the major process reengineering or staff training recommended by other marketing automation vendors.

This is clearly a successful argument: Act-On just added its 350th client, and is on pace to triple its total over the next twelve months. Most buyers are small businesses, and many are in financial services, insurance, retail, healthcare, and education rather than high-tech. Installations at large companies are often used by small groups who want an alternative to the existing corporate marketing automation system. The system’s $500 per month starting price and no-annual-contract policy encourage such casual implementations by reducing the financial risk.

In fact, the risk is so low that Act-On has actually reduced its free trial period from 30 days to 14 days and eschewed the freemium offers used by competitors Genius and LoopFuse. Raghavan argues that marketing automation systems are too complicated for freemiums to work well. (For what it’s worth, Genius and LoopFuse seem satisfied with their freemium results. Beyond inducing trial, freemiums can also attract long-term users who will move up to the paying version when their needs expand. Agencies who use freemium versions at small clients may also use the same product for larger clients who will pay for the system.)

If it sounds like my discussions with Act-On have centered more on strategy than product features, that’s correct. Raghavan argues marketing automation vendors' success is now determined less by product features than repeatable, profitable processes for sales, deployment, and support. He said Act-On has developed these processes and can now safely accelerate its growth with predictable results. One change the company plans is to expand its managed services, such as help with deploying campaigns. Some of these services will come from Act-On and others by agency partners.

Although Act-On sells the product as a replacement for existing point solutions, the company says clients do use additional features fairly quickly. The most popular are drip marketing, anonymous visitor identification, sales integration, and Web traffic analysis. Complex nurture campaigns and lead scoring are used much less. Under-utilization of advanced features is common across all marketing automation systems, but it may also reflect Act-On’s sales pitch that clients don't need extensive changes to their marketing processes. Similarly, it's tempting to argue that clients' demand for managed services reflects the company’s deploy-first, change-later approach, but that’s also something that clients of all vendors consistently request.

In short, Act-On is doing quite nicely by ignoring the industry conventional wisdom that process change and planning are necessary for marketing automation success. But that doesn’t mean the conventional wisdom is wrong. After all, Act-On’s main sales proposition isn't about marketing automation: it's about making existing tasks easier. It also helps that Act-On is selling to small, consumer-oriented firms, whose marketing automation needs are genuinely simpler than at larger, B2B marketers. In fact, although Act-On is selling to slightly larger companies than micro-business specialists Infusionsoft and OfficeAutoPilot, its resembles them in several ways.

The real test of Act-On comes after clients have gained experience with it and find they do need marketing automation training and process change. So long as Act-On has the features and services to support this – which I believe it does -- the clients should be happy and renew. But if they find they’ve outgrown Act-On because they need more change more than they originally realized, they’ll have to look elsewhere.