Slashdot Mirror
Organizing Your Web Services Division?
"For example, at the credit union, the position I was hired into was brand new. They wanted to bring their web site in house. Their solution was to hire a Manager of Internet Development (me) who was responsible for determining the needs of the credit union, setting up the servers, doing the development and programming, and maintaining the site. No staff, and they wanted the site up as quickly as possible. I spent most of my time reporting back and forth between the VP of marketing and the Director of IT. When they finally figured that wasn't going to work and tried to have me report to one department, they couldn't figure out which one it should be so they eliminated the position and outsourced the web site again.
I am running into the same thing at the county. I came on about a year ago to a web site in shambles. The previous 'web team' consisted of an Internet Administrator, a team leader, a webmaster, a web data specialist, and a web temp. The team leader wanted them to be their own section, but unfortunately, he did it by power-playing and burning bridges. The Director of IT came through and broke the team apart, firing the team leader and the web data specialist, releasing the temp, and splitting the remaining team between the Distributed Processing Management (DPM) and the Network Administration sections. The other webmaster left about two months after I came on, leaving me as the sole webmaster for 3 sites of around 80-100 thousand webpages. We are finally back up to staff (another webmaster and a web-data specialist). The challenge we are running into is that in order for items to get on the site, they are designed by the departments, approved through our communications department, then passed on to us to integrate into the site. If we have a server problem, we have to contact Network Administration, even if it is something like having a Data Source Name set up.
To further challenge matters, the manager we report to has 28 people who directly report to him, including us.
With the size of the sites being what they are, it wouldn't take much for the whole thing to fall apart, and I am trying desperately to prevent that from happening. I envision an Information Architecture being put into place which would allow us to work on content management, instead of building these pages by hand. But I seem to run into obstacles every where I turn."
dublab will make you feel it.
Isn't this why you get paid?
I mean come on! If you're not competent enough to do your job, why should anyone here tell you?
Oh, and by the way, FP beeyatch.
Avoid committees's like the plague. No matter the organization, too much dependance on a seperate group making the decisions will only foster a failure. Its not as much a problem with splitting the group into sections, but rather the people that oversee those groups. can you make descions together between sections, or will you be needing to go to the higher up, which then calls the dreaded meeting.
If you are working at a startup, you can do it all
yourself (hardware, design, code, maintenance, etc).
The age-old conflict is the IT people want it
maintainable, always up, and conservatively
designed, marketing wants to do things on the
seat of their pants without advance notice..
I separate the server maintenance from the updates. I manage a colo, server, backups, and the cgi parts of the server, the contractor of the week does the design & updates. The tools I have built are all designed to have no ongoing maintenance from me (IT reporting).
If you can make that clear from the outset,
you can co-exist well with a marketing department
or a PR branch etc that needs an effective
publishing platform. These boundaries sometimes
result in conflict:
Do it quick
Do it stable/well
but rarely does it become catastrophic if you
work with good people.
We have found that having a small web services team that is directly accountable to a high level manager is important. In fact, our web site is currently down because of a reorganization in that area.
We had similar problems in the past few years. In fact, our web services used to be run by the IS department, who used our web site to try out "new and cool" technologies. That is one reason why we were, until recently, running web servers on HURD, BeOS, and AtheOS platforms, with little coordination amongst them. We also had trouble keeping up to date on patches, and nobody seemed to want to learn enough about the novelty OSs to support them properly. In effect, our web site became a toy for bored administrators to tweak. Not a good idea when we've got millions of dollars in funding riding on the public and the Congress being able to measure our progress as an organization.
So, in a business, it would probably make sense to have a dedicated VP who oversees the web site, along with several senior technical people who approve changes. Although everybody hates red tape, it's simply not a good idea to trust a couple of recent CS grads not to mess up the company's image by goofing up the web site. Changes should all require approval, and unapproved changes should be grounds for dismissal. That is how we are doing it, so stay tuned to see how well it works when the site comes back up...
~wally
The best models I have seen have different groups responsible for the design, the coding, and the hardware. The design people are usually the freaky Mac types, the administrative people are the moles who are really into that kind of thing, and the software people are the attractive psychologically balanced ones who do the actual work.
Ahem. Excuse me.
Seriously, the separation of labor between these three camps works out best because it allows each group to maximize their specialty. If you have some designers with good HTML skills, then your coders can, for example, just drop in, custom JSP tags where appropriate without having to mess with the web server or design principles. A group consisting of people who have a lot of knowledge in one of these areas and a little knowledge in each of the other two tend to perform best. I hate to use the word "synergy" but it really is appropriate here.
Depending on your resources there are other areas to consider as well. Q&A is extremely important and can help the developers to more efficiently debug. Content writers and proofreaders are important as well; someone who can tell the difference between "your" and "you're" can be a real boon to your professionalism.
But the basic web team, IMHO, should minimally consist of the three core elements I listed above. The most successful projects I have worked on have been variotions on this theme.
- Rev.Spend some time building a plan and make damn sure you have a single reporting line. The dotted-line/solid-line/dual-reporting scenarios are just a great way to get yourself used as a pawn in a power game.
How independent the team is all depends on how integrated the web site will be with county operations. Is it generally just a static window on the county or can folks actually lookup info, search jobs, read government reports, etc, etc, etc?
I'd say the web team should be an independent entity of the IT department. The web team would be responsible for the servers, security, processes, and base content. They would also be responsible for the style-guide (and the gloriously fun task of getting it fully approved).
Each county department/division gets their own standard 5-page setup (mission, history, news, blah, blah). Additional content is reviewed and billed back to the request department to keep from overextending the IT/web group with the demands of each department. Since each department typically has their own budget lines (and cost centers) this might work well for you.
Speaking of which, who has the largest amount of data on a web site? Could it be IBM, or C-Net, or e-bay....?
Hence three different departments.
First the content... Pure Sales / Marketing
Second the machines... Operations
Third the Software... IT
The webmaster should NEVER be the operator. But could be the a coder - but not the best use of their time.
You want the:
Webmaster to be artful, help place a "good face" on the company.
Operator make it run day in an day, plan for backup and outages, fall overs and upgrades.
Software to build it on the machine and make it look like the webmaster wants. Coding in the main frames, the sql servers and the web hosts.
You need to think of web development as program development. GUI=Webmaster, Business Rules=Software, Data Storage=Operator. With that model work gets done, by the right people for the greatest impact.
Maybe you could help make the criminal search system a little nicer :-) We utilize it at work (we do employment screening) and the wenches who use it always complain about its uptime.
Easy. Fire them all.
Good luck to you and your dealings with the red tape of public employment. I have been there.
the deacon...that's all you need to know for now
Clear Channel Communications' web site department (which maintains over 1100 sites) is a completely seperate entity and it works excellent.
Web people need more B&D, imho. Nothing like good Fagan Inspection to keep things sane and controlled.
So I decided to change things around. I formed a committee (yes, an evil committee, but bear with me). The members were mid-level people from each of the dozen or so internal constituencies. We met twice a month to review the direction of the site. They would inform me if something I was doing bothered them, or if they wanted something new added.
I would also meet individually with these members, and we'd prioritize the content for their department. I'd then collect all of the content "wish lists" together into one master list. The committee members could then see the entire list, and see where their items were on the list. I'd have the Chief of Staff review the list as well, and since they knew he'd signed of on it, they couldn't hassle me when a priority of theirs wasn't high on the list.
The best thing was that these mid-level people were the sole content managers within their departments. All content for that section of the site had to go through them, and then on to me. If they slacked off on the job, or didn't help me out enough, I could talk to their supervisor about it. If they did a good job, they got kudos from the boss, if they screwed up, everyone in that department knew why their part of the site was lagging.
It wasn't a perfect solution, and it can't be applied in every organization, but it helped me maintain my sanity as the sole person developing and maintaining a 1,200+ page site. Even though I left three years ago, the structure I put in place is still being used, and the site has more than doubled in size.
The key to it all is that you have to try and understand human nature. Everyone wants to sharpshoot you if you fail, but by setting up a system that puts other people in the loop while still giving you primary control over development, you can keep your sanity, and more importantly, build a system that will work for everyone.
Best of luck - I feel your pain.
Read the EFF's Fair Use FAQ
I really don't know what you mean with your "suck your duck!"
Could you explain please? I really want to know it.
I don't know what the optimal is, but this is how it works (more or less) at my company. I will not give out any individual names nor the company I work for to protect the guilty:
Marketing owns the content on our web site (the www.com address). They can access and update the docs themselves.
A separate entity, E-Business, owns a separate site which runs our app. The people in this org manage the application and content on this server.
The IT department ties everything together by managing all of the server hardware, os and network configuration.
Basically, this structure helps our company by keeping the people who know a certain area best working in just that area. The coders do the coding, the designers do the content, and the IT guys manage the hardware aspects. Most of the time this works out ok.
The frustrations come in mostly when Marketing or E-Business run into network problems, and have to turn to the IT department. Unfortunately this particular IT department isn't a very reliable one, often ignoring or conveniently 'forgetting' about a problem until you have to pester them to look at it. Problems that take 15-20 minutes to fix are left broken for days until you come back to bring it up again. And again. And no, they are not overworked. They also have a habit of changing things around (IP addresses, router configuration, firewall 'upgrades') without notifying the other departments who run dependant services. If these guys worked for me, they would have been fired long ago for their work ethic and lack of organization. Since they don't work for me, and they're not even in the same branch in our org hierarchy, I can't do that. You have to make sure you've got a decent IT manager who can control this sort of thing. Too bad this particular one can't (his employees reflect his own behaviour).
You can accomplish anything you set your mind to. The impossible just takes a little longer.
I experience a similar phenomenon at the non-profit I work at. There is no Web department. Nor even a real Web budget. Our 'Web Master' is in the Communications Department. I'm the 'Web Advocate' in our Advocacy Department. Our Publications Director does his own coding. Our IT Director has a consultant that tweaks our server box - which is now in-house since our previous host merged and lost what little brain power they had left. There is no central managing authority or oversight and instead we have an 'IT Committee' that meets every other week to try and do the job of a Web Director. The budget is all ad-hoc and barely gets us through. We also manage content in six languages, relying mainly on volunteers and interns. I argue for a Web Department and a budget, but org. management just doesn't want to hear it. Needless to say there is nothing resembling a strategy or plan. It's a constant battle just to get through the week.
That's life in the real world. Get used to it.
I have worked on large web sites where my teams were almost all cross-functional, dual-reports with other departments and on teams where I had direct control of all my assets. Both can work, but the problems are a little different on each side.
When working on cross functional teams, the key is politics and bridge building. Find someone in each key department who is going to be fully reponsible for your contact and work within that department. This sounds self-evident, but I have been places where random PR people seem to take whatever parts of various web projects that interest them. This doesn't work.
Once you have a single point of contact, constantly keep that person's manager in the loop on projects and goals for web efforts. Give them some mindshare in your projects. Get them vested in what you are accomplishing for them. Also carve out very defined work boundaries and proceedures for the duties of your contact person.
Next, when something works, make sure you slather praise though the organization that hits the managers of your cross-reports. You team will get its due because the stuff is obviously up on the web site. Make all your praise to the people you need to keep happy to keep getting work from your cross-departmental memebers.
I have found this to work in organizations with the right culture. The thing is that with cross-departmental teams, the majority of your job becomes consensus building and coordination. I have had this eat up almost all the time I would rather be using on real work. But it is the nature of the buisness.
In a perfect world, I want my own team, broken into servers, content production and coding support. In my experience this produces more consistency and quality in my team's projects and tends to ensure more uptime.
But the problem here is that a smaller number of managers have a vested interest in the success of the web team. This can hurt in budget fights. And the budget fights WILL come.
The other problem here is on the content side. Coming from a content background to start out, I always want direct control of my content folk. But a couple of dedicated content people, while frequently better and more professional content producers, aren't as close to some of the things going on in a large organization and things that could or should be in the content pipeline can get overlooked that way.
I know this is unfortnately just a post dealing with the edges of the ugly politics, but in larger organizations, most of the reality of leading web dev teams seems to come down to dealing with the political issues that are instrumental to being able to get the work done. A sad thing. Maybe I have just had bad luck selecting jobs.
7. What we cannot speak about we must pass over in silence.
I work for a large company doing web development for the external site. There are several problems with our website. Since our group (Internet Engineering) is in charge of future development for the site, we've charted out the problems and their solutions. Here is a short synopsis:
Where I work is not important; what's important is what you can learn from our mistakes. Every major company has organizational problems with their web site. Your company must take these issues and deal with them now as opposed to later. A content management system is an invaluable tool in helping with these issues. Most feature workflow routing so you can have one manager signing off on several people's projects before they get published. You can then also hire graphic designers without having to hire a complementary Dreamweaver jockey; good systems create HTML and correct menus for you in a lot of cases. Most take the Developers / Graphic Designers / Managers / Administrators approach, where they place you in one group and you get different tasks according to what group you're in. This may or may not work for you; the good thing is that in a good content management system, you can customize it to fit your needs.
It's great that you've asked this question when your group is still small. It shows a lot of forethought that the older, larger companies didn't have time for. In a lot of respects, you're lucky: you can design an ideal system. Just make sure it will scale and that you can easily upgrade if some new time-saving feature comes out in a year or two.
I generally agree with your graphics, coders, and hardware divisioning.
However, one issue is whether or not the coders should be part of the rest of the IT group, or just web-dedicated. As the web matures (or changes), there will not be a strong distinction between web coders and other coders IMO. It only exists now because a decent web GUI standard has yet to emerge (like SCGUI, plug plug).
(begin quote) If you have some designers with good HTML skills, then your coders can, for example, just drop in, custom JSP tags where appropriate without having to mess with the web server or design principles. (end quote)
Sometimes it should be the other way around IMO. The code/database controls the general layout, but indidivual "contentlets" with graphics etc., are inserted by the graphics people via an interface created for them. Which strategy one takes perhaps depends on the nature of the site. If the graphics must be different, changing, and "cool" because you are selling jeans to teenage girls, then perhaps graphics should be the outer layer. However, if providing timely information is more important than visuals, then I would let the code/database control the outer layer.
Table-ized A.I.
At my company we have a large online content services group. They are responsible for the development of the content as well as publishing it. The leader of the content group is at the same level as other IT groups. The content group however has no control over the servers themselves. The servers are controled by the systems administrators as they should be. Just as the content group maintains the content and makes the rules up for how things will look and flow, the sys-admins control the systems updates and change controls used on the servers. This works well for us, because our teams work together to make things happen. Content publishing is a part of IT. If you can't get results by having your teams work together, you have other political problems that won't go away very easy. I used to work for a local goverment and I would bet the later of here is true. Sorry but there isn't any way around those dang politics.
Integration in an organization is key. Depending on how "modern" the management in the company is, an idea you might want to try is a participatory, cross-functional team.
Try having people from different areas (Finance, Marketing, Human Resources, Operations, etc.) with different levels of authority coming together to work as a part of the team. This will give your department the decision-making power it needs in the organization, while maintaining a broad perspective in terms of everything you will need to successfully set up and run the site.
Yeah, it makes sense to have a single group in charge of the company Web site. This eliminates duplication of effort and -- more importantly -- makes sure your whole site has uniform navigation and graphics so that users don't give up in disgust. Stuff like a single sign-on is also important for huge sites.
For a medium-sized site, the team should consist of:
It doesn't have to be exactly 5 people, but some number of people who combine to cover all these skills.
Oh, yeah, I forgot a really important one:
This last person isn't necessarily more important than the others, but he/she is harder to find. Lots of projects fail for lack of this guy, the one who tells the other suits "That would be too expensive," and "That would take too long," who sets expectations so that when you (the programmer) have done a damn fine job, the rest of the company knows it.
If you haven't been in industry yet, you have no idea how important that is.
Andrew Wiles
a**n + b**n != c**n for n > 2
Most organizations that I've worked for use a generic staging/production server system where new content and services are first pushed to staging for testing/QA, then pushed to production. This has numerous problems, and we've had many situations where the push to production led to bugs visible to the world.
I proposed to my company that we move to a see saw system. We, in effect, create two sets of servers, each of which alternate between production/staging.
When we have new services and content to deliver, we do the following:
1) Perform an exact copy from the production to staging servers.
2) Setup the staging servers to reflect the changes via VSS GetLatest.
3) When the staging looks A-OK, the staging servers themselves are switched to become the production servers. The production servers are taken offline and become the new staging servers.
Things have been wonderful since this change. First, the switch is more instantaneous because we only have to run a swap script on our load balancers and a few other servers. Second, it allows hardware replacements to occur on a timely basis without interrupting service.
After taking my licks in this industry for a decade, I think I've learned some hard lessons about management:
1. You need a person who makes the decisions. PERIOD. Anyone who tries to bring issues above his/her head should be slapped down by his/her seniors.
2. This figure needs to diligently listen to the needs of his organization. However, this figure also needs to resolve conflicts quickly and decisively. If this means hurting someone's feelings, hurt them (and maybe sweet talk them later). If this means firing someone's sorry ass, fire them.
3. Experience and "the right way" takes priority over personal feelings and political correctness. Encourage controlled conflict. Stir things up at designated meetings. And if any level of yelling or bad feeling arises, there should be a mandatory drink-beer-and-talk-shit after work.
This is how things get done.
In our office the main web administration is handled by IT. IT is in turn considered part of Operations. Operations consists of HR, Facilities, Food Services, Library Services, and IT. Basically all non-financial, non-engineering, and non-marketing.
Is to not do something silly like Slashdotting your own sites. You said you were the senior webmaster, and yet you put out URLs to be Slashdotted?! Either you like pain, or you just want to be called in on a Saturday (which is another way to describe pain, I guess).
;-)
If all you have are silver bullets, everything looks like a werewolf.
Well, At my previous job we had the best web team I've ever seen as far as efficient website management. Let me break it down for you.
The company had about 200 total employees, our web team was about 5 people, and then there was the Database team & the content team, both which we worked closely with. It was extremely important that each of the 5 people knew a little of everything and was an expert in one important skill.
1). Team leader - very good with security, and FreeBSD (or server OS of choice). And general server setup & maintnence.
2). Network guy - had a nack for network setup & design. Worked with the foundry routers and junipers, as well as helped with server maintnence and shell scripting little things.
3). Perl Scripter - wrote many scripts to help automate and monitor servers. Helped with server administration and setup etc.
4). Jack of all trades - proficient in most areas and able to help out where things were needed. Excellent to have.
5). Jr. Sysadmin - the lackey you get to make patch cables and drill in the rack. train him on
server setup & maintnence etc.
Our team was responsible for the setup/maintnence & overall design of the website. We setup load balancers, content management (not the actuall content), like seting up jserv, monitoring the servers, setting up the Sun boxen for the database etc.
The database team just focused on the data organization etc. and they worked with the Java (content) team to produce the dynamic pages. We did the build process and installed everything into our network layout, they didnt have to worry about anything but whether their content was correct (and looked cool).
I think this design is very efficient, because it allows each team to have a very specific focus and to not be distracted with understanding some other areas that are close, but not related.
Our main problem (which you really want to avoid) was that there was a big gap between US and our Manager (Director of Engineering). Basically we just didnt have enough communication and there was alot of distance between the teams and him. Basically fell apart at that point.
If you have mostly static pages, then you might be able to get out from under the IT group by moving the site to a web hosting provider. IT might even welcome that. But if the site is well integrated into many databases maintained by the IT group, then you need them, and have to live with multiple reports.
In our organization the Webmaster reports directly to the President, but she has consistently moved the site from static pages to database generated ones (mostly created from databases kept for internal administration). As that trend continues, IT (me) gets more infulence. But I never try to tell what to do with a static page.
I honestly doubt that "Walter Bell", a systems engineer at NASA would have time to post so much on slashdot.
I work for a government contractor, and have seen government systems, the idea that they would tinker on niche OSs is completely laughable. They are usually the LAST people to move to the latest platforms. The computer technology used by gov't is hardly cutting edge.
So while this guy might have a web server with HURD running, it definitely is not commmon to see.
My experience tells me that separating information and function is essential and should be the first thing to do.
One of the more elegant approaches I've used here is having near 100% of the data on the site stored in a database (or databases).
(Performance is gained by good use of caching components in the application code)
Since you didn't describe what kind of functionality the "company" requires it's a bit more tricky to give a proposal on the technical side. However, going with a MVC (Model-View-Control) architecture will be fine regardless.
Application logic is run from and AS (Application Server) which is based upon templates for presentation. The templates could for be for ex. JSP pages using custom tags for creating the design of the pages. The data provided to the pages would be retrieved from the database by business objects and delivered (wrapped in data structure beans) to the JSP templates.
That way, you could have the designers focusing on creating templates, the programmers taking care of the business logic (and creating custom tags) and the database people making sure that the database is running at an optimal.
When it comes to updating information, I'd suggest spending some time creating a content mgmt. tool (not very hard), or let an outside company do it.
With this CMT "ordinary" employees could change the data without having to involve the IT department.
As for the organization structure regarding the web aspect, I'd suggest having each organization within the company appoint one responsible person, who is the one channelling and deciding what information will be allowed to be published on the site. This person can of couse appoint some other people whom have this authorization as well, but that is not for your IT group to decide or even bother with.
The focal points within each org. can send you an official request for people who should have the ability publish info and all you do is add these people as authorized users in the CMT application.
This way your group can focus on what you are good at while the other orgs. can do what they do best.
When a decision making person in an org. decide that some information is to be allowed to be published onto the site, they can enter the date and time for when the information goes live and press the "approve" button. The rest will be handled automatically.
As you see, this requires no specific restructuring of the company. Instead you can probably continue using a structure which the company has probably been using for years (and which is hopefully already working well).
I think it's important to realize that by utilizing the Web, companies do not have to radically change their way of working when it comes to publishing information. What was earlier published in press, ads, brochures etc. can be done much in the same way. The only difference is that a CMT tool is used instead of sending the info to the print house.
If you do not think your company will be able to create this kind of solution (provided this makes sense to you), I could probably get you in contact with one which have. In that case just drop me a line at: chris_7d0h.antijunk@yahoo.com
(Note remove the [.antijunk] spam protection from the address)
In a society that believes in nothing, fear becomes the only agenda ~ Bill Durodié
Heh, I've seen (participated in?) the IT vs. Web development wars and it's not a good idea to have Web as a client of IT because IT is usually pissed/jealous/scared/disgusted at/of/at the Web group for many of the same reasons that the Web group can't stand IT.
.001 rev of some plugin or why we woudln't run machines with no security, hand out root passwords, etc.
I was on the IT side, and we had a shitpile of varied systems and thousands of end user computers to keep running *securely* and in whatever harmony the software would allow, within our budget, and within the complex politics that is most corporations. A complicated job by itself.
The web people I worked with were usually nice, I even respected some of them. But they didn't get the big picture of the total environment as well as we (tried to) do, and they were always at the bleeding edge of everything and were pissed because we wouldn't do massive client-side upgrades every time there was a
I'd agree with you that they need to be seperate, in fact I'd vote for totally seperate -- like in another building across town. I think better web sites (better interop, better functionality, etc) comes from not even dreaming about having any influence or control over the desktop environment.
If you have a VP who is so stupid he/she can't figure out this stuff for you then you are going to have an impossible challenge. The problem is that if you take some of the advice I have seen on this thread (create your own well thought out plan and present it blah blah) which are great ideas you run the serious risk of becoming the enemy of the VP who realizing they are failing horribly will start to attack people below them. Its human nature. So screw trying to teach this idiot what they should already know and would either only take credit for or sabotage. If the CEO/president of your company is so lame they can't hire the right people which really is about the most important part of their job then I say colelct your checks, keep quite, and leave ASAP. And of course warn the rest of us who these idiots are so we can avoid working for them or demand outragous salaries for taking on the stress of leading from the bottom. All that talking will do (if upper management doesn't see the problem already) is get you into trouble. -m
We have a three managers. One takes care of all the application development/graphics stuff, one for the DBAs, and then the last manages everyone. We have about 10 Java developers, and half are contractors. We have 6 or 7 DBAs in house, but only half know what they are doing. Also, we have a full staff of business engineers (RUP, IEEE SCM) but they are a joke, and produce zero deliverables.
The last project I worked on, I did all the req engineering, research, arch, coding, testing, documentation, html, graphics, DBA stuff, and the project was a complete success. No kidding!
I don't know. I guess I am saying that I am one of those people that sometimes tries to do everything on his own and usually does a good job. I am sure this can be bad sometimes, but I don't care. Try to find more people like me that work hard and aren't afraid to cross over boundaries with skills they don't possess. find people that place great stake in research and learning, and aren't going to say, "that's not my responsibility." This is a proven benefit of BPR (business process engineering) techniques. You'll always get that lamer that argues that she can't learn everything because she wants to specialize in one thing. I say, stop being a pussy and hit the books; your mind can handle an immense amount of complexity.
The first step is to try to get it into one department - working together with others is fine, but you do need the final authority on the website and anything you need for the site(s).
Seeing as how that's problematic, your best bet is to nudge it in the direction you want it to go in small steps, and not in one big shove.
Make sure you have the people you need. At the least one server wiz, one perl wiz, one graphics/content wiz and someone who can make them work as a team.
The place I now work at (a whopping 2 employees, yes yes, startup) already has this on paper - except that I'm doing all 4 of those things right now. In the future (read: when we make money) there will be people hired to focus on specific needs - thankfully I managed to get a sensible structure going from the get-go.
There is no sig...
This is a touchy subject for most companies, everyone wants to play politics since websites are high-visibility.
I've seen it at companies where the web team was part of IT, a separate group, and once where they reported to marketing. There are pros and cons to each of them, it really depends on what kind of company you're with and how strong the group leaders are.
IMHO things work really well if the IT group is *competent* and has a good director/manager and you're part of that group. You get fast response for your needs on the IT side, and protection from marketing's wild ideas (let's go with an all flash website!). It will work best though if you have a good, strong IT manager who will listen to you, stand up for your team, and make sure you get the support you need. This is good for a corporation because marketing can dicatate the content, you can manage the web technology and IT can support you and the systems.
The separate group can work, but really only if you're in a small company or if you have a high-level boss. You're going to be stuck between IT not wanting you to change things, and marketing's wild ideas. Usually marketing has a high-level VP close to the CEO, unless you have someone big enough to fight for you you'll be at their mercy.
Which is just as bad as being under marketing IMHO. Marketing usually doesn't have a clue about technology, and they come up with wild ideas about how fast things can be done or random toying with tech just "because".
Again, it really depends on your corporate structure and the people involved. Personally I think you can forget about crap like VP of Web Services. I can't believe companies waste their time with this crap. You know how much $$$ I've seen on VP of Janitorial Engineering, VP of Food Services, Chief Privacy Officer, etc. etc. Most companies need less time-wasting VP's and Chief's and more Indians getting things done.
Webservices will enable corporations to turn the faucet on in gaining valuable information. It will commoditize information and make it a free flowing resource. I think you will know that web services have come when you can order a web service like you order electricity, telephone, or water service to your home. You just call up and they turn it on. That said, the question is really what we geeks need to do to stay on top of this. After all, I don't see many electrical, telephone, or water technician walking around with $100K pay checks. In turn, the IT profession will meet great changes. I came across an interesting personals site recently. Imagine this site, GroovyBooty, turning its database of pictures into a web service. What are the implications for that? Well, first off, online dating will be forever changed. Postings from GroovyBooty could be syndicated around the world and we will see a massive shift away from owning the information to distributing the information. Distribution will become the key, rather than information creation. That said, the real question becomes ... what do WE do? If www.GroovyBooty.com becomes a mere database, how does it GroovyBooty.com continue making excess profits?
I'll leave that up to the sales and marketing guys. Heheh!
Remember that there were those of us doing real work on Cisco and SunOS that paid the bills, including all the money that the web group squandered, back when internet meant the home office LAN and all the field offices linked together, not some dot-bomb fairyland.
I don't doubt that there are real pikers in many IT departments, but in the one I work in, there are very few. We were, however, subjected to a rotating collection of child "web designers" who knew fuckall outside of VB script and could have cared less about anything that they couldn't put into their Palm Pilots in under 30 seconds.
As for security, about half their sites (hosted externally) were owned in a week long period. Even after that they wondered why wouldn't let 'em run NT boxes open to the world.
First off, let me say that of all the responses, you need to be listening to corky's response the most. Having worked on some huge sites myself, I can confirm that he appears to actually speak from experience. As for my experience, I've worked at 4 companies since I got into the Web in 1994. Here's the breakdown:
In my experience, the best solution is autonomy. Being in Marketing, you are going to butt up against people who are driving technology decisions but have no clue what they're talking about. Marketing people should not be managing developers. In IT, the Web will either be too rigidly controlled, or will be a hodge-podge of all the various new technologies the geeks wanted to play with.
At SST, the saving grace has been the content management system, which we've built entirely ourselves. It's simply a few Web forms that dump data into a database. And then other Web pages use that data for display. But what is sooooo important is that each form has an owner, and that person is responsible for filling it out when needed. Each time we deploy a form, we free IT from the manual labor of building and rebuilding pages upon request. We are also able to tie in automated signoffs -- some content goes to a manager for approval, some content is published immediately. Whatever the case, you need a content management system. A small, nimble one.
My Greasemonkey scripts for Digg &
>I currently work for a county government as their senior webmaster
... issues
... quickly
> I ran into
> No staff
> they wanted the site up
> web site in shambles...
> sole webmaster for 3 sites...
> challenge we are running into...
> size of the sites...
> whole thing [could] fall apart...
> desperately...
> obstacles every where I turn...
Hey, cool! You've got a job!
What was the question?
Unfortunately, it seems as though your contract job hasn't given you the experience it takes to realize that different agencies in the government actually have different setups, needs, and "corporate" cultures.
Your post is tantamount to saying that because the Navy standardized on WinNT, it's obvious that nobody at NASA runs Linux.
Please check your arrogance at the door before you post again.
~wally
A company that has its web team reporting to marketing just screams "we don't under the internet." Marketing executives simply don't tend to have sufficient experience with engineering and administration issues to understand the goals, challenges and advantages of having a web site.
A lot of companies see their web site primarily as a marketing tool. That may be, but running a web site is completely different than laying out a catalogue or brochure.
- Scott
Scott Stevenson
Tree House Ideas
The first thing you have to ask yourself is what is the main function of the website.
If the site is used for brochureware then it is a marketing function. Marketing should have somebody to oversee the website, with both technical and content people working on the site.
If the site provides a service, then it falls under operations because it is part of the day-to-day business of the organization.
I have a half-half situation in my organization. I have a brochureware corporate site and a very complex intranet application that is web-based. The solution:
1. The outside website is run by marketing and sales (it comes out of their budget, so it is theirs). Whenever they need programmers to work in it they buy our time with their own budget.
2. The intranet is run by operations. We assign internal resources to deal with it.
3. The outside website has some automated components that are being fed from the intranet. These dynamic components are maintained by operations. Any and all content issues are solved by Marketing and Sales.
Final recommendation: Keep the team as small as you can afford. Avoid yourself heartbreak later. I already see you will need a project manager, a good programmer and somebody to deal with content. The graphic artist you do not need full time (unless you are one of those lucky bastards that has a good programmer that also has decent graphic skills!).
Pedro
----
The Insomniac Coder
I work in Sili[y]con[trived] Valley and we recently put out an ad for a "Content Production Assistant" and received over 1800 resumes the first day ( a truly scary story in its own right). While hidden in a cube not far from the people plowing through applications I heard "Ooooooohhhhhhh this one is a Weeeeeebbbbbbmaaaaaaasteeeeer!!" followed by gales of laughter. That and the associated comments that were yelled out make me think these poor souls weren't faring too well. This has led me to wonder what the new catchy titles are? Content Czar? Yes - this is unabashedly off topic - mark it thus.
http://www.archive.org
Over 100 Terabytes and growing FAST.
View snapshots of websites over time from 1996 - present. Very neat!!
My background has included working in very small groups that grew into very large groups servicing very large sites (two million+ hits per day, profitably), with a couple of different sites now, both still profitable, so I know that what I'm going to describe *has* worked, but of course that doesn't necessarily mean it'll work for you.
As a small team we focused our responsibilities such that they were broken up into design/implementation and operations/administration. As we grew, we broke design and implementation into development of new software and upkeep of existing software (bug fixes, integrating new pieces - very small team by comparison, very high importance though).
One of the reasons for breaking things up along these lines is that these three groups have both very different priorities and very different economies to be gained in large scale.
For example, for a development team, the priority is often getting a design or piece of software completed and ready for release. Efficiency, testing, stability are all frequently secondary, as are the skills and resources associated with those motives (perhaps unfortunately). This team develops skills in the areas of rapid application development and pulling together product needs from other departments. They frequently get the bulk of the "fame".
For an upkeep team, stability is the primary goal. Such a goal is frequently directly at odds with speedy product development, and also often at odds with redesigning to increase efficiency. This team gets paid to be careful (and fired if it breaks too often) and manage release standards and versions. This team also has to be careful not to be perceived as a bottleneck for progress, and they're sort of the middle ground between "development" and "operations", so it's very, very important for them to maintain very good relationships with both of those teams or they won't exist very long, since management tends towards forgetting their value. As harsh as it sounds, another reason for seperating out this team from primary development is that as an IT manager it gives you someone to point to and fire when there's a major screw-up and heads must roll -- without severely impacting your larger development team and ongoing projects, since this team is only charged with post-development.
An operations team isn't really focused on the internals of a product at all - instead their goals must always be to provide services and infrastructure to the other departments and manage costs.
For example, the operations team frequently provides network management, system backups, and hardware allocation. This also entails standardizing and documenting system and software builds; it includes performance tuning, which requires gathering and archiving data related to performance of all infrastructure, from CPU utilization to licensing to network bandwidth.
The operations / development seperation is a particularly important way to benefit from economies of scale. For example, network monitoring and management is something that can have a big cost in terms of people time and software costs - it hardly ever makes sense to have multiple teams using multiple systems. Instead it's usually much simpler to have a team that provides 24x7 network monitoring using Spong, Netsaint, Openview and that ilk. Ditto with licensing management, which can be a huge problem and a big money-waster if ignored (as well as a big liability). This team doesn't need to be highly skilled software coders, although it does help if they can do *some* programming to automate systems software. They frequently don't get noticed unless they deliver a big cost savings on a regular basis and they have to be careful to make sure management knows they did so. Operations needs a very good manager to keep track of what they do and how much they cost the company versus how much they save the company or they risk being considered dead weight no matter how effective they are.
Just my 2 cents. (And hey, where the hell did my cent-sign key go???) Hope it helps.
t @ wondious.com
I noticed you have an email form on your website. It reminded me about something almost no webmasters seem to take responsibility for. Email customer service.
MAKE SURE SOMEONE IS READING AND RESPONDING TO EMAIL.
You wouldn't (well, you shouldn't) have incoming mail or phonecalls that no one checks and responds to, so don't do it with email. If someone goes on leave, either monitor their inbox, or set up an autoresponder (always check they are actually working, and keep checking it).
If you have a large volume of email correspondance, have a seperate email response group, don't allocate it as something to be done in spare time.
Have a procedure about how all this is to be done, and how long it should take for email to be responded to.
There will be people who are supposed to follow this, who have no idea how email really works.
Email is not an optional extra if it's used for public or client contact.
If I rang a store or office and left a message, and got a return call 2 weeks later, I wouldn't be impressed. Same goes for email.
Think this isn't the responsibility of a webmaster? who knows the mail links are available to the public?
Unless there is a notice near all email links that it is not a priority for replies, or to expect a delay, most people will assume a quick response.
Now while the monitoring etc may not be what the webmaster is doing, someone needs to make sure the rest of the organisation is using the website properly.
Letting the web group know when email addresses become defunct, or staff leave/arrive.
Letting the web group know when staff are going on leave, or are away for the day (if one day makes a difference in your work).
A website is like a virtual office in a lot of ways, and it all needs to function properly.
Don't be a weak link in the chain.
Hi. A coworker of mine sent this to me and I was thrilled to see discussion of this problem, since I've been stuck in this same sort of place (which is where????) for over four years. My first experience as a web designer was as a contractor for a large state agency. I was there for about three years and over time the situation as far as our autonomy / ability to get stuff done got better AND worse. On the one hand, we managed to evolve into a fairly autonomous team where each member had a job. There were three web designers / programmers, one content reviewer / agency liaison, a group supervisor (who also did lots of liaison and content work), and then there was a boss, somewhere out there, who only got consulted when we had problems or something major. We did a lot of work with individual departments within the agency as well as people in the state counties/regions. The worst problems we faced were that 1) we had no one "upstairs" who really had a clue how to fully utilize us, which was bad in terms of getting priority in budget, 2) we had hardly any connection to IT, so it would take MONTHS to get simple stuff done, and 3) the site ended up being really high maintenance because we had so much material coming in from the departments we'd buttered up to work with us. These three problems came together in a pretty nasty way right around the time I managed to get a new job. Basically the boss got an idea that we should go to a content management system so we could handle all the maintenance and content that was coming in. This is not a good thing to attempt when you don't have a lot of $$$, when you have very limited staffing as far as people who *directly* support you. Our agency had IT people for the whole agency, but we didn't really have our own IT people or if we did it changed like every week and they were forever playing catch up. In the end, the contract on the content management system got bogged down in red tape (complicated grants, funding, etc.), the administration sort of forgot about us, and finally most of us jumped ship becuase we were all sick of it. I landed with another gov't group, an IT department within a technical research division at a large university that's basically all technology, engineering, etc. Our group works a lot better than the one at the state ever did, but for the web site, they're still not ready. Basically they have me (working in the IT group which mostly does Java programming, plus our intranet) doing the whole web site myself, working with someone in another dept., the head of communications. We in turn work with various department/laboratory heads. She has her boss, I have mine. It's been good and bad but lately bad because the other woman, my partner, has had to go to half-time and it leaves me in the lurch, since she routes all content. The other bad thing is that 90% of our web site being eliminated due to 9/11 security concerns. I think I just sant the heck out of gov't jobs, steady as they are, but then I think there are nasty problems and dumb work set-ups everywhere. Wiebke
That's how we ended up where we are now. :)
keep it up :o)