0 Autonomy Promises to Automate Delivery of Tailored Marketing Messages

Summary: Autonomy is a leader in enterprise search and content management (it owns Interwoven). Its concept of "Meaning Based Marketing" comes quite close to my idea of "content grazing" as a way to reduce marketing complexity while delivering the right content to each customer or prospect.

In my last post, I proposed the (somewhat tongue-in-cheek) term of “content grazing” to describe automatically extracting small bits of information from company documents and feeding them to prospects and customers. The notion had been on my mind for some time, prompted by a sense that traditional approaches to content creation and distribution are fundamentally too expensive to deploy in full. That is, tailoring large content streams for buyer segments and funnel stages is just too much work for both marketers (who can't afford to create the content and campaigns) and buyers (who don’t have time to read the results).

The good folks at Autonomy apparently had similar thoughts. Back in July 2009 they published a paper Meaning Based Marketing that describes using a collection of Autonomy technologies “to truly understand your customers—drawing on everything from transaction history to cross-channel interactions, user generated content, customer and community behavior, as well as third party content—and act on that knowledge to deliver the best performing, most accurate, and relevant content to each individual visitor. Automatically.” That pretty much says it all.

The Autonomy brief has a slightly different scope from my own concept, perhaps because it’s tailored to fit the Autonomy product portfolio. For example, it includes Web content archiving, which I wouldn’t have considered, and excludes creating content from relevant pieces within a larger document. I do consider the latter quite important, because it’s a key to reducing the content creation cost and in making the information more digestible for customers. But the core components of meaning extraction and self-adjusting content selection are certainly present in Autonomy's description.

Specifically, the paper outlines a fully automated process of:

- analyzing unstructured content to infer the views, needs and preferences of customers who create or read it

- generating customer profiles based on the contents and associated behaviors

- uncovering customer segments within the data

- delivering a tailored customer experience (that is, content recommendations) based on its understanding of both the customer and the contents

- optimizing interactions through real-time multivariate testing

The list also includes market analysis, based on sentiments and trends, as well as the previously-mentioned content archiving. Both are useful additions to the vision.

The paper doesn’t go into much technical detail beyong mentioning that Autonomy unifies the data through its “Intelligent Data Operating Layer” (IDOL). But I suspect you could find much more information elsewhere on their site: Autonomy itself doesn’t seem to be conserving its content-creation budget.

Nor, come to think of it, did I notice any particular personalization or intelligent targeting on my own return visits to their Web site. Either they’re very good at this stuff – and know I’m not a real sales prospect – or they haven’t quite gotten around to deploying these services for themselves. Either way, the concept remains valid and it’s good to see that someone is working to make it happen.

0 2010 Will Bring New Features to Demand Generation Systems

Summary: the demand generation market will continue to grow in 2010, and it may attract some new, big competitors from outside the industry. But the real excitement will be features that expand the scope of demand generation products to support inbound marketing, better measurement, and more efficient content creation.

2009 was a year of tremendous growth for demand generation systems (a.k.a. business-to-business marketing automation. By some measures, it's looking more mature: buyers are appearing outside the initial niche of software and technology companies; core functionality is well understood and largely consistent across products; vendors are expanding scope to include new users at existing accounts (in particular, sales departments); pricing is under pressure; and companies are starting to specialize in different customer segments.

On the other hand, there are still plenty of new entrants; few pioneering vendors have failed or consolidated; and related software vendors (in this case, CRM, email and Web site management systems) haven’t yet introduced me-too products. Perhaps most important, many potential buyers still don’t understand the value provided by these systems—although vendors are working very hard to educate them. So, on balance, I'd say the industry is still in a fairly early stage: late adolescence, if you will.

What will 2010 bring? Continued sales growth, for sure: that’s easy enough when you’re starting with a small base. We can also be confident that the feature trends I described in my review of 2009 will continue: better support for social media, greater access for sales departments, and more flexible reporting. I do expect vendors to converge on more standard social media features. These will probably combine the content-sharing and activity-tracking capabilities that different vendors now deliver separately.

There’s also a reasonable chance – although this prediction is less certain – that sales access features will blossom into deeper cooperation between marketing and sales in managing prospect relationships. There's no question in my mind that such cooperation will appear: it's inevitable as marketing’s role expands beyond lead generation to long-term relationship management. What I don’t know is how quickly this will happen or whether the sales access tools will be the connection point. One reason they might not is that sales access tools are used by individual sales people, while broad marketing and sales integration is likely to be controlled by senior sales management.

So much for the rear view mirror. Here are some predictions that are larger departures from the immediate past.

- me-too products. It's just a matter of time before CRM vendors (yes, I mean Salesforce.com) and Web content management vendors decide to compete seriously for marketing automation business. Frankly, this is so obvious that I'm almost embarrassed to mention it. But I wouldn't want anyone to say I failed to see it coming.

- inbound marketing. The work of generating Web traffic through search engine optimization, paid Web ads and expanded Web content has so far been performed outside of most demand generation systems. These are important marketing activities and they are a natural extension of demand generation systems, even though they require closer integration with (or replacement of ) Web content management and Web analytics. Note that Webinars and social media, which are also inbound marketing devices, are already being added to marketing automation products.

This type of extension—supporting new tasks for current users—is typical of maturing products once the core functions widely available. It also implies that vendors specializing in these areas will add their own marketing automation features to compete. HubSpot particularly comes to mind, which is a testament to their own marketing skills.

- external data. Many demand generation systems already make it easy to look up data about prospects from sources like Hoovers or JigSaw and to infer the location and company of anonymous visitors from their IP address. Certainly those features will continue to grow. But there’s another trend that's very pronounced in the consumer marketing space, which is using consumer panels and surveys to measure responses that aren’t captured within the company’s own systems. I haven’t seen much analogous activity among business marketers, but think that will change as the technique becomes more common and as business marketers accept that internal data will never provide all the answers they really need for effective marketing measurement. The task for the marketing automation vendors is making it easier to integrate such data and, in cases such as ad-embedded surveys, to generate it.

- content grazing. I'll explain that label in a moment. The idea is to squeeze the most value from existing marketing content, rather than creating new content for each project and situation. This implies two complementary tasks: being able to extract and classify nuggets of information from existing marketing documents, and being able to deliver exactly the right nugget in each situation.

The underlying insight is that there’s so much information available today that people don’t have time to digest large blocks of it. Rather, they want be fed bite-sized chunks that meet their immediate needs. Hence, the term "content grazing": it's like eating appetizers instead of a full meal.

Today’s marketing best practice is the opposite of content grazing: it’s to develop many different campaigns that deliver large volumes of content for different situations. This is expensive and it's exactly what prospects don't want. The alternative is automated systems that extract and classify content from existing materials, including many such as blog posts that would be created for other purposes. Other automated systems would can select and deliver the correct content during each interaction.

Basically this is the challenge of simulating a human conversation. It’s possible that some solutions will be based on automated customer service agents already used for other interactions. I haven’t seen this applied in a marketing automation context, but suspect it’s a path that marketers will be forced to explore as they recognize the full cost of conventional content-heavy approaches, and that buyers don't want them anyway.

0 Marketing Automation System Trends: What We Found in the Raab Guide

Summary: Social media and access for sales people were the two big trends among demand generation vendors last year. But enhanced reporting was the most common improvement of all. Could marketers finally be ready to spend on measurement?

I’m pleased to report that the 2010 edition of the Raab Guide to Demand Generation Systems is officially available today, with updated entries on all vendors (alphabetically: Eloqua, Manticore Technology, Market2Lead, Marketbright, Marketo, Neolane and Silverpop Engage B2B).

Preparing the updates gave me a good review of where developers focused their efforts last year. Even though this is limited to the vendors in the Guide, it's a pretty representative sample of the industry as a whole. Here’s a quick look at what I found.

- Social media. At least three vendors (Eloqua, Marketo and Silverpop) introduced new features aimed at improving marketers’ ability to use social media. What’s most interesting is that no standard approach has yet emerged. Eloqua focused on making it easier to embed sharable links within conventional marketing assets. Marketo added features to capture Twitter posts and Helpstream customer support interactions within a lead’s activity history. Silverpop made it easier to add social media handles to lead records so these could be used to send messages.

- Sales access. The same three vendors also added new tools to give salespeople better access to information the marketing automation system has captured about their leads. But in contrast to social media applications, the sales access modules were remarkably similar. All aimed at showing the activities of selected leads, typically by showing overviews and trends, and then letting users drill into details. The vendors also charged additional per-user fees for these modules. This contrasts with traditional demand generation pricing on database size and/or activity volume, but is the way sales automation systems like Salesforce.com are usually sold. These modules open up a major new revenue stream for the demand generation systems while simultaneously giving sales departments a greater reason to support marketing's purchase of the systems. Even though the modules clearly trespasses on the CRM vendors turf – inviting a potentially devastating counter-invasion – the opportunity seems irresistible.

- Upgraded reporting. You already knew that vendors were adding features for social media and sales access, but did you realize that nearly everyone (five of my seven vendors: Eloqua, Market2Lead, Marketo, Manticore Technology and Marketbright) also made substantial improvements in their business intelligence and reporting capabilities? Popular new features included user-customizable dashboards, better user-defined reports and more extensive standard reports. I take this as evidence that marketers are demanding more sophisticated reporting from their vendors, and suspect further improvements are on the way.

- New user interfaces. Market2Lead, Manticore Technology and Silverpop all introduced major interface upgrades. The focus was less on adding new capabilities than on making existing functions more accessible. I don’t need to remind you that usability is a critical point of competition among industry vendors. But as older vendors revamp their interfaces, it will become harder for buyers to differentiate along those lines. This might lead vendors to highlight the structural differences in their campaign engines, which are ultimately more important for usability than the visual interface. But, the structural issues are much harder for buyers to grasp, so this might not be an effective marketing approach. Could this lead vendors to compete on other grounds entirely?

- Anonymous user look-up and data enhancement. At least three vendors added or enhanced features to use IP address to identify the company of anonymous Web visitors, and/or to look up prospect names and other data about those companies in directories such as Jigsaw and Hoovers. I won’t name those three because the other vendors may have similar capabilities. In fact, anonymous visitor identification and enhancement have become pretty much standard features: today, it would be an exceptional vendor who did NOT make them available. These features also tie into both social media and sales access modules. They illustrate how the role of marketing has changed from simply gathering leads and handing them to sales, to building and managing prospect relationships.

So much for 2009. Many of these trends will surely continue in 2010, but I think we can expect some new directions as well. I'll talk about those in my next post.

0 LoopFuse Offers No-Frills Demand Generation

Summary: LoopFuse offers attractive but limited demand generation functions at an easy-to-swallow price.

It’s been nearly a year since I took my first close look at the LoopFuse OneView demand generation system. I didn't write about them because the vendor was planning some major improvements and it made more sense to publish a review after these in place. We reconnected in December and the product is now ready for its close up.

The resulting picture is much prettier than before because the main changes in LoopFuse have been improvements in the user interface. It is now an attractive system with wizards to drive major functions and extensive in-line help messages to clarify next steps. The help functions are built with Helpstream customer support technology, which also provides a knowledge base and online community.

The process to set up multi-step campaigns (which the vendor calls “lead flows”) is now quite straightforward. Users first create the lead flow and link it to recipient and suppression lists that will determine who enters the flow. They then define the flow itself by following prompts to add nodes for activities, conditional decisions, waiting periods and retrial periods. (A retrial resembled a waiting period except that the system keeps retesting the previous condition. For example, the retrial node might check every hour to see whether someone has responded to a previous email. Come to think of it, it would probably make more sense to build the retesting into the conditional node itself.)

Esthetics aside, however, the actual capabilities of the lead flows are still somewhat limited. The only direct action available in an activity node is to send an email; all other options involve updates to the CRM system such as adding a lead, changing data, or assigning an activity. This leaves out other tasks that I consider basic, such as updating data within the LoopFuse database, removing the lead from the flow, or adding the lead to a list. (In the case of data changes, the omission is intentional: Loopfuse argues that such changes should be made in the CRM system and replicated into Loopfuse).

The conditional nodes make up some of the deficit. For example, failure to meet the node condition can remove the lead from the flow. (An activity to remove leads from a different flow will be available around next April.) The conditional nodes can check for several specific conditions, such as email responses, Web page visits, CRM lead status, data values and lead scores. These probably serve most purposes, but some users may be frustrated by the inability to combine several conditions within a single node or to specify more than two branches as outcomes.

Leads enter LoopFuse lead flows when they join the associated recipient list. Lists can include leads with specified data values on the lead record, that originate from a particular Web site or form, or that visited a particular Web page. Several conditions can be combined and lists can be static or regularly updated. But lists can't be selected based on lead scores, email response or activity levels. Again, processes that depended on these would need to incorporate them through conditional nodes in the lead flows.

Lead scores themselves can be based on data values in the lead record, email response or Web pages visited. However, the system cannot base scores on activity patterns such as "three Web site visits in the past week". The system stores one score per lead. Scores are recalculated every hour, which does not support immediate reaction to score-changing events.

LoopFuse provides a graphic email designer that can generate both text and HTML versions. Emails can be personalized with data from the lead record, including the assigned salesperson if available. But data from account or opportunity tables is not available, even though it’s imported from CRM. Nor does LoopFuse support any type of A/B testing, either in the email definition or its lead flows. Each email is tied directly to just one email “campaign”, although the email campaigns themselves can be reused in multiple lead flows.

Unlike most demand generation vendors, LoopFuse does not host landing pages or Web forms for its clients. Instead, the vendor provides a wizard that reads existing, externally hosted forms and generates modified versions that will post data into the LoopFuse database. Another wizard helps users to generate HTML forms from scratch. Either way, the new forms must be copied into pages hosted outside of LoopFuse.

LoopFuse will also provide tags that can be placed on other pages on a client’s Web site to track and report on Web page visits. It lets it generate some page-oriented Web analytics reports similar to Google Analytics.

Reporting is another area that LoopFuse has significantly strengthened in the past year. In addition to Web analytics, it provides detailed reports on Web site visitors, including reports that link anonymous visitors to their company through their IP address. The system also tracks lead movement through sales funnel based on a combination of its own data and information imported from CRM. Other reports show results from email campaigns and lead capture forms.

LoopFuse also gives salespeople a report showing recent activities by their assigned leads, allowing them to drill into each lead for details. A “company dashboard” report lets both sales people and marketers see all visitors from a particular company, again based on IP address. The report shows both known and anonymous visitors and lets users connect directly with external databases including Hoovers, Jigsaw and Zoominfo to look up additional information. LoopFuse can also alert salespeople by email when high-priority accounts perform specified activities captured in the system.

LoopFuse does a particularly good job with CRM integration. Because its sales process depends heavily on free trials, the company has developed a self-service wizard that guides users through the process of connecting to Salesforce.com, including an automated check for whether LoopFuse has been granted access permissions in the client’s Salesforce.com installation. (Apparently this is a very common omission.) Users have field-by-field control over how data conflicts between the systems are resolved. In addition to Salesforce.com, the vendor has a standard integration for SugarCRM.

Pricing in LoopFuse is based on email and Web page volume, with no limit on the number of names in the system database. This is unusual but not unique: Pardot and LeadLife take a similar approach. (See my recent list of demand generation vendors for an overview of competitive pricing.) Rates are quite aggressive: $750 per month for 50,000 emails and page views combined, or $1,250 per month for a much more generous 250,000 total.

LoopFuse was founded in 2007 although it was largely in stealth mode through early 2009. The company currently has more than 50 paid clients.

0 Spredfast Offers Systematic Management for Social Media Campaigns

Summary: Social Agency’s Spredfast helps marketers schedule social media campaigns the same way they schedule paid advertising. Cool.

It seems like common courtesy to listen to an existing conversation before jumping in with a comment. If social media worked the same way, companies would first buy a monitoring system to track what’s being said, followed by tools to respond to comments made by others. Only later would they initiate conversations and, eventually, provide tools to help their friends spread the word.

Silly me. I should have known that marketers talk first and listen later.

I’m not talking about personal style, although the decibel level at any marketing conference speaks for itself. But a recent eMarketer article B2B Marketers to Increase Social Spend cited two surveys that showed this is also a matter of policy.

Specifically, a study from Visible Technologies and SiriusDecisions found the most common use for social media was to “generate awareness” (25%), while another study in B2B Magazine found the top use for social networks was “thought leadership” (60%). True listening ranked fifth in the Visible Technologies/SiriusDecision survey (“monitor and respond” at 14%), and third in the B2B Magazine survey (“customer feedback” at 46%).

On reflection, this makes sense. Marketers are primarily interested in getting out their messages. Perhaps this is an old habit that will change in a customer-driven world. But I suspect that marketing results will always be driven by activity, and results are ultimately what matter. So I’ve now revised my expected sequence of social media activities to start with broadcast, only then followed by monitor, respond, initiate individual conversations and empower advocates.

Armed with this insight, I was much less surprised when Kenneth Cho of social marketing agency Social Agency told me that his new social media campaign tool Spredfast had been purchased immediately after release by major companies including AOL, IBM, HP, Cisco and Porter Novelli. Although I’ve seen plenty of “listening platforms” like Radian6, Alterian Techrigy and Scout Labs, I hadn't previously seen a system aimed primarily at managing outbound social messages. (Now that I'm looking, though, I find that ObjectiveMarketer seems to offer something similar.)

Of course, the listening platforms can also post social media messages, as can the social media features now found in many marketing automation systems. What distinguishes Spredfast is that marketers can schedule their posts through the life of a social media campaign, rather than simply replying or initiating conversations on demand. Spredfast supports on-demand posting too.

Another key feature is that Spredfast supports multiple “voices” of actual or constructed individuals, each having accounts in multiple channels (Facebook, Twitter, blogs, etc.). The campaign calendar lays out of scheduled events by all voices over time, and is color-coded to show whether a particular event has already been delivered, is ready to go, or still needs approved content. This looks strikingly similar to the media plan for a flight of broadcast ads and serves very much the same purpose.

Users can drill into an event to add the content itself including bit.ly links that allow the system to track the click-through. One particularly nice feature is that when users assign the same content to multiple events, the system will automatically create different bit.ly links for each event. This makes it easy to track results for each event independently.

Spredfast's developers also recognized that large companies will have many different people working on different aspects of a project. Users can be assigned rights to specific campaigns, voices and events, with precise control over who can view, edit and approve content. Pricing is based on the number of campaigns, not users, so large organizations can incorporate as many people as needed.

As the bit.ly tags suggest, Spredfast also pays substantial attention to measurement. It provides three major summary metrics:

- activity (how much content the system is publishing),
- reach (the number of views, friends, followers, subscribers, etc.), and
- engagement (numbers of comments, retweets, likes, etc.).

Top-level reports summarize these by campaign and let users drill down to see detailed statistics by channel and voice and, ultimately, the actual content such as comments or reviews.

The system archives content and responses so they remain available even after they are dropped from the social media platforms that originally carried them. In addition to cumulative statistics, Spredfast displays daily statistics for the past seven days, giving a sense of trends.

Perhaps wisely, Spredfast's developers drew the line at reporting the raw numbers for its metrics. Users who want more elaborate scoring, perhaps applying different weights to different kinds of activities, can export the raw data and calculate outside the system. Similarly, Spredfast makes no attempt at relating social media programs to business results such as leads or revenue.

The system does provide what Cho called a “minimalist” listening platform, which can automatically search across public listening tools (Google, Google Blog Search, Social Mention, Twitter, Boardreader, Bing) for key words, and present any results so users can review and republish or reply to them. It also provides an RSS reader for feeds selected by the user, as well as a site indexer that can show the frequency of different terms in user-specified blogs as a word cloud. This helps users tailor their posts to encourage coverage.

Spredfast began its public beta in mid-November. The system is a vendor-hosted service. It is available in a free version with limited functionality; a $50 per month standard version with one campaign, no collaboration and no metrics after the first month; and $250 per month enterprise version supporting all features for up to three campaigns.