If that's the core objective, it involves several skill sets. The more of them that can be (truthfully) represented in a single contact by the customer, the better for the customer, because it reduces the amount of time that the customer has to spend on non-core activities. For example, if you are a carpet supplier and fitter, you need to care more about carpets than HTML and the vagaries of which search engines use what algorithms - what you need to care about is whether this marketing activity through this agency brings in sufficient profit and builds sufficient brand equity that the activity should be sustained, grown or decreased, or whether the problem is the agency.
I believe that at least the following skills may be required to make a web site effective - the only question to my mind is the decision as to where to draw the line(s) on SEO:
- Domain Name & Administration
- Email accounts and mailing list management
- Web and mail server administration (headers, etc)
- Web page templating (download speed, accessibility, usability)
- Information Architecture
- Visual design
- Legal framework (US and Europe have fundamentally different approaches)
- Ethics, privacy policies, oversight and administration
- Directories, Banner Ads, Sponsorship, Interaction with Portals, Vortals, etc.
- Integration with other marketing communications (telesales & other offline activities)
- Metrics (abandonment, conversion, visitor sources and referrals, search phrases, etc)
- PPC, PFI and other SEM
- Buyer Behaviour Models
- (Sometimes) e-Commerce systems
- Marketing Information Systems
- Web application development
- Workflow - content management, order management, issue tracking
- Content Management Systems, and their specific features and failures
- Inventory control and presentation
Part of the definition problem in this thread has been, I think, that different customers have different needs, can offer different complementary skills and that success measures for some sites are only measurable over a whole company's activities.
We've definitely done work for customers who do not understand that the web is not a magazine or a TV and we've spent time educating them about what they can do and how this might affect their entire business activity, if they had suitable product and services. We've dissuaded some people from spending quite large sums, because, in our opinion, we didn't think they'd be satisfied with the response (the buyer model didn't involve a phase in which any internet activity would be significantly useful and there were no other stakeholders in sufficient numbers that the cost-per-contact would be better than picking up the phone, and the brand would not be sufficiently enhanced by a large and complex website). We've also spent time with people who've solidly understood their market and have taught us about their customers and what their marketing needs are - they just can't translate that into technology.
FWIW, I've come from a tech heavy background, and I am co-founder of several small successful businesses. I do understand the business processes, the marketing activities and the separation of sales, marketing and customer service (and other components). I'm also taking a part time course to get accreditation with the UK's Chartered Institute of Marketing - and despite hanging around marketing for more than a decade, the courses are still eye-opening. We work with some major agencies for large accounts where we take on technical/marketing specialist roles and we work directly for smaller businesses where we are their online agency and provide everything, down to the hosting, in order to be able to offer search engine marketing and optimisation.
One hat isn't enough. One definition won't work. Just as there is not a single definition of marketing that works for every expert, or even a single definition of CRM. Being effective involves flexibility and matching your services to customer need - and if your services are limited, selecting only those customers that you can satisfactorily service.