Seven Details That Bust Web Analytics Budgets

In the early days of Web analytics, deployments were relatively straightforward: you bought the software, bought the hardware, got your support contract and designated a few people to maintain the system and analyze the reports. That was pretty much that. Those days are long, long gone and definitely not returning. In fact, as complicated as cost/pricing is today, it’s only going to get more complex as vendors add more services and partner arrangements.

This article takes a detailed look at licensing choices, analytics options, marketing tool integration, services needs, training and technical support, long-term hardware needs, and human resource tasks and roles. The fine print can come back to haunt you, so each and every one of these areas should be thoroughly reviewed and detailed in your budget. As in any technology deployment, if you fail to plan, you plan to fail, so let's take a closer look.

1. Licensing Choices

There are two models available for conducting Web analytics: Software as a Service (SaaS) and licensing. In the SaaS approach, the vendor hosts the servers and software used for site measurement. This may also be known as an Application Service Provider (ASP) model. All upgrades to the basic software are typically included in the subscription fee. Pricing is based on a subscription for a set contract length, usually one to three years, and you can typically avoid big up-front costs in favor of paying a fixed, monthly fee over time.

In the conventional, licensed software approach, you purchase the servers and software used for site measurement and everything is located and managed within your enterprise. Pricing can be based on a “perpetual license,” which has no end date, or terms can be reviewed on an annual basis. If you own a license, you may need to pay for major new releases to ensure that you are working with a supported product. Subscription fees and license fees are often determined by the number of page views your site generates per year or per month. In SaaS scenarios, some vendors may provide a base number of page views within the initial subscription and then increase your fee on an incremental basis if you go into higher page-view tiers. In license scenarios, vendors that review your traffic on an annual basis may also raise your license fee.

To develop this initial estimate, vendors will either ask you to provide a traffic estimate or sample server log files. It's likely that you will filter certain items from your site prior to data analysis, so make sure you're estimate doesn't include some of these extraneous resources, such as images, frames, spiders, bots, and traffic from domains you might want to exclude, such as that of your own organization.

The numbers of servers and domains to be tracked will often have an impact on your subscription and license fee as well. This is an area to watch closely, especially if you plan to track hundreds of distributed sites --a scenario common among media companies, enterprises with local affiliates and outlets, and organizations that run multiple intranets and extranets.

SaaS customers will also need to determine whether there are costs related to data storage. Most vendors will store data for at least 13 months; some for the lifetime of the contract. Some vendors store raw data; some only the reports generated. Be sure you understand the pricing and storage format options available to you. Obviously, the more data you store and the longer you store it, the higher your costs. If you export the data back to your servers, then storage and query becomes your long-term expense.

2. Optional Modules

So, how do the vendors put together those great demos with all of the cool reports? Well, one of the big reasons demos are so dazzling is that they are using their entire product suite and can show off all the functionality available. Keep in mind that if you want all that functionality, you may have to pay for it. These additions to a core product are often called "modules" or "extensions," and they add functionality to the core product. Let’s look at a few examples:

Reporting. Subscription-based products may come with limits on the number of reports, filters and dashboard report options. Some vendors charge additional fees for using Excel-based reporting options.

Data analysis. Beyond the standard, "out-of-the-box" reporting offered, there are often options to report on raw, unaggregated data collected before the information is summarized into reports. Access to this data often comes at a premium. Some vendors also offer advanced segmentation tools (typically at an additional cost) that support complex segmentation scenarios. When reviewing vendor offerings, it's important to itemize the features, understand the costs, and decide what you really need to satisfy your analytic requirements.

3. Marketing Tool Integration

One trend gaining traction is the ability to marry analytic data with marketing tools, such as keyword bid management, e-mail campaigns, and internal search or cross-sell engines. This eases both decision support and the creation of business-rule-driven campaigns. Vendors typically feed analytic data to marketing tools through both internal solutions and partner relationships. You should consider these integrations before you purchase and have a sense of what you will require. Costs depend on project scope and whether implementation support comes from your analytics vendor, e-marketing supplier or in-house resources.

4. Professional Services

All vendors offer consulting services. The options start with packaged implementation consulting services to help you with your initial installation, whether you are using SaaS or licensed software. These services may include some measure of requirements analysis, data collection strategy, cookie deployment strategy, page-tag-variable customization, page-tag implementation, and data verification.

It makes sense to review vendor implementation services because it can help jump start the process and help you avoid mistakes and delays. Many users of analytic tools regret taking the “do it yourself” route to save money only to find themselves trying to fix problems and delaying implementation. It's critical, however, that you understand what the vendor provides in their implementation package. You'll often discover standard, premium, and customized options, so make sure you get what you want and only pay for what you need.

In the area of data integration, it's likely that you will want to do one or all of the following at some point:

Export Web analytic data to your enterprise data warehouse

Import offline data into your Web analytic database

Export Web analytic data to marketing tools (See “Marketing Tool Integration”)

Most vendors offer APIs that provide the blueprint for enabling data import and export. However, if you don’t have the resources to implement the API, you may need the vendor’s professional services teams to complete these tasks.

You may also require professional services to develop customized reports, provide guidance on the collection of challenging data, such as Web 2.0 activity, share best practices regarding the placement of tags, and database and server optimization if you are hosting a software license. You should obtain database and server optimization insights before finalizing your purchase decision, as it will figure in to your budget allocation and your choice of vendor. In using any professional services, be clear on your requirements and ask for a written statement that provides ample detail and clarity on what you'll be receiving as deliverables as well as the cost of those services.

5. Training and Technical Support

Vendors may offer myriad training options (on-site and off-site, standard or custom, paid or free Webinars) for all user levels (administrator, power user, report user). Sometimes training is included with SaaS subscriptions. You may also be able to use training hours as a bargaining point in your negotiations with the vendor on total costs as some vendors may use training as a “loss leader.” It's always a good idea to get training on basic operation of your analytics package and administration training so you have staff who can train others. Web analytics practitioners repeatedly tell us: don't scrimp on this expense.

Once your system (or service) is up and running, vendors typically offer a range of support options. Some of these are bundled with your service while others cost extra. The following services are typically bundled at now extra cost:

Account Manager

Online Knowledge Base

Online User Forums

Online Help and Webinars

Online Documentation

The availability of telephone and e-mail support may vary with the level of support you purchase. For example, basic support may provide business-hour support local to the vendor, whereas premium support may provide 24/7 access. To ensure the level of support you require, some vendors offer Service Level Agreements (SLAs). Note, you may need to set up SLAs with your internal technical teams if you are relying on support to maintain a licensed analytics tool or if you are "borrowing" resources to implement any tagging or data integration development.

Dedicated account managers are becoming a rarity in the software world in general, but as in other categories, Web analytics account managers are typically incented to "upsell" you on new products and services.

6. Hardware

Hardware costs are only a factor if you are hosting your own solution (or importing and storing large amounts of data from a SaaS provider). If this applied to you, consider server capacity, server software, disk drives and disk arrays, tape back up, and data storage costs. Reporting on large datasets can be extremely CPU- and memory-intensive. Data collection is I/O-intensive. Depending on your configuration, collection and reporting could have network capacity implications.

Infrastructure costs can be capitalized over the life of the analytics initiative, but keep in mind that you may need to purchase additional hardware if your system starts to explode with data. This could be challenging to forecast and must be managed efficiently, especially with log-file-based systems. If you don’t come up with a solid data expiration and archiving strategy, the glut of information could bring the analytics application to a standstill. Analytics vendors like to point out that storage is cheap today, but the people needed to maintain those boxes are not. This brings us to our final topic…

7. Human Resources

All too often we forget that it's people who run the show, not the technology. If you forget about the people side of Web analytics, your initiative will fail (see "Roles" depiction at right). Web analytics is not necessarily a high-maintenance technology, but it is high attention. Let’s look at the roles and tasks typically associated with implementing, maintaining and exploiting your Web analytics application:

Marketing / Communications Champion

Senior-level executive with profit-and-loss responsibility for enterprise marketing initiatives Sets marketing strategy, based on analytics input

Project Manager

Maintains and manages analytics vendor relationship

Manages analytics team and technical support resources

Manages Web site measurement projects to include initial implementation and new initiatives

Acts as primary interface between analytics team and senior management

Manages user education and training

Analyst

Develops metrics requirements with stakeholders

Determines data required to develop metrics

Analyzes reports, calculates metrics

Provides strategic and tactical recommendations based on metrics analysis

Website Manager

Manages Web content, marketing and visitor acquisition programs

Manages relationships with all Web channel stakeholders

Supervises Web marketing, content and technology teams

Ensures suitable analytics tags or mark-up are applied to all templates and pages Developer

Performs variety of customization and integration tasks required for data collection, extension, import, export and so forth

Database Administrator

Participates in data integration

Performs database maintenance and optimization tasks associated with software-based Web analytic applications, such as archiving, data expiration and tape backups

Server Administrator (for licensed installations)

Maintains servers, installing security patches and updates

Monitors data collection, data processing, and imports into database

Monitors software and hardware components of site measurement system

Application Administrator

Sets system permissions for those who will have access to reports

Schedules and publishes reports

Provides diagnostic skills and the first level of support on system problems; point of contact to vendor support.

These roles and tasks may require further splitting out or may be combined, based on the scope of your Web analytics implementation and the size and breadth of your organization. In other words, each role could be part of a job, an entire job or it could require multiple specialists. Larger enterprises have multiple Web site managers. In smaller operations, the same person may serve as Developer and Application Administrator, as well as Server and Database Administrator. The work doesn't go away, however, so you have to account for the tasks. When setting your budget, consider how and whether these tasks can be accomplished with current resources. Will you need vendor support, additional staff or consultant/contractor services? As in every aspect of Web analytics deployments discussed in this article, you'll set more realistic goals and develop a more realistic budget if you plan for the need.

This article is based on an excerpt from the Web Analytics Report, which evaluates 13 major Web analytics vendors. Reprinted with permission from CMS Watch (www.cmswatch.com/Analytics/).

Phil Kemelor is Vice President of Strategic Consulting Services for Semphonic, a Web analytics consultancy. He is also a lead analyst with CMS Watch and consults with enterprises looking to select Web analytics technologies or apply best practices.