About PM4Web

The PM4Web blog was born as an outlet to return knowledge back to the web development community. My goal is to share my experiences as a project manager from over the years in a manner which helps you succeed with your own projects.

19 November, 2010

10 Tips for Keeping Momentum on Web Projects

Tricks for avoiding stagnation and block points on a web project.

There are many hidden perils lurking in the midst of a web project. So many little devils which conspire against you delivering on time. The nature of a project is that it is a unique undertaking, so unless you can see into the future, chances are there will be hiccups along the way.

If we know there is trouble on the horizon, then we can be prepared. The idea isn't just to have a plan ready for dealing with what maladies may befall your project, but also to use strategies to nip things in the bud before they become a problem.

To that end, I present 10 techniques I use to keep things moving along smoothly on a project. Some of these may seem counter-intuitive at first, so you will have to use your best judgment about what approaches you feel comfortable using.

Before I get into the meat of the article, I should say I'm not covering obvious things like 'use a project schedule' and 'write a spec' - if you aren't already doing that, then no amount of clever tricks will save you.

1. Get the ball rolling early on things involving third parties - a good example is live hosting, you can get seriously bogged down if you leave this until just prior to launch. If you don't have full control over the server it makes you heavily reliant on tech support to do things for you. Another example is e-commerce gateway integration. Do you really want to wait until the last minute to find out if your code is talking to the gateway provider's server properly?

2. Have settings accessible via the CMS - I've found it's preferable to have settings adjustable via the CMS rather than have them in code or in a string table. Options like Google Analytics script code could be placed inside a System Settings or Website Configuration page. Doing this lets non-technical people deal with last minute changes.

3. When there's two ways to code a simple feature, code them both rather than wait for approval - if the features each take only 15 minutes to implement, just code them both and enable one. You have a 50/50 chance of getting it right. If the client decides they'd rather have it the other way around, you switch on the alternate option (preferably via a setting in the CMS). The amount of effort you would expend going back and forth to get a concrete decision from the client is probably going to exceed the 15 minutes it would take to just code the alternate feature.

4. Put in the wrong content - this may sound counter-intuitive and you'd expect it to lead to an unhappy customer every time (but it doesn't, most of the time). This is only for preview/staging versions of the client's project, not their live site. If the client hasn't made content available to you, then use the closest match. Something off their old website perhaps, or workable place-holder text. This does one of two things; the client may say "hey, that's the wrong text", to which you reply "OK, can you give me the correct text please?", or it gets them started by providing a basis to work from (as opposed to starting from scratch).

5. Have a change request budget - this concept is straight out of PRINCE2 methodology. It basically says when you make a fee proposal/tender, include a chunk of budget for unexpected features. So for an eighty hour project, I would have between 5-10 hours of change request budget. The good thing about this is when the client asks for something off-spec, you don't have to wait for approval or try and convince them to pay for it (they already have). It can also help cut down on admin tasks since you don't have to track the item for inclusion in a future invoice.

6. Get clients thinking about what they hadn't thought about, early - a good example of this would be a social media strategy utilising Facebook (assuming it's appropriate for the business). A client may not have thought about Facebook, bringing it up early may have them saying "hey, I hadn't thought of that". That gives you time to provide them with some advice or reading material. Leaving it to the last minute may throw them off, they may even decide to delay launch to consider social media options.

7. Don't just offer options, offer a recommendation - if you talk to a client about a feature where there's two or more paths to choose from, first describe the available options, then the pros and cons for each, and finally make a recommendation on what's most suitable and why. This helps clients who are cautious of making a decision because they don't understand the technology at play. An example would be if a client asked whether its better to integrate a blog into their website or use a well-established one like Blogger.

8. Use a bug tracking system - scope creep can seriously side-track a project, potentially delaying the original launch date considerably. An issue management system lets you deal with off-spec features in a controlled way. If a client asks for a feature and you say no, that's just bad customer service. If you say yes to everything, you'll be coding forever. Instead, if you say "yes, but its been flagged for post launch implementation", you should be able to segment off features for future coding. Unfortunately this doesn't always work, sometimes clients are adamant that a feature needs to be included immediately.

9. Start early on tasks which require client approval - an example of this would be a storyboard for a Flash animation. The client may need a few days to review the document. Connected to this tip is starting early on things which only the client can do. A good example would be applying for an Internet merchant account. The bank will most likely only deal directly with the client due to security considerations.

10. Begin some work before the project deposit arrives - if you're fairly certain the project is going to proceed, do a couple of hours work even before the deposit has arrived. This would be stuff like preparing the project folder, checking that the domain delegation details the client has provided actually work, minor things (but no coding). This gives you a head-start on the project, its creating momentum from the get-go.

Dilbert - project plan for 10 second task

Keeping momentum on a project isn't just about following documented procedures. The role of creativity and a willingness to take risks shouldn't be underestimated. Novel solutions which produce results are either praised or go unnoticed, whilst failures may attract negative feedback. The trick is for your successes to significantly outweigh your failures.

Join RSS FeedSubscribe to RSS Feed.

06 November, 2010

Understanding Google Analytics Reports

A primer on interpreting the statistics inside analytics PDF reports.

This article will help you figure out what the statistics inside a Google Analytics report mean. This is not to be confused with the full service provided when logged into the Google Analytics website. From within the Analytics website, you can set it to automatically email you a periodic summary report (e.g. once a week). The report itself arrives in your inbox as a PDF attachment.

One last note before we move onto the interesting stuff; this article has been written to help non-technical people (e.g. an accountant who's just got their hands on their first website). So, if you are a computer science major, you may read parts of this article and think to yourself "that's not entirely true". And you would be correct, but I have intentionally made a few generalizations for the sake of brevity.

If we are going to talk about Analytics stats, we need to first ask ourselves "what's the point of all this data?" Sure, the figures can demonstrate that a ROI has been achieved (e.g. "was my money well spent, or should I have just gone down to the casino instead?"). But information without application is largely useless. The statistics should guide and shape the decisions you make in future (we'll cover some examples in a moment).

Absolute Unique Visitors
Arguably the most important figure is contained on page 2 of the report; Absolute Unique Visitors.

This stat is largely self explanatory; it's the number of people coming to your website. For most people it's the number which says whether their website is being successful or not. When a website first launches this figure may be quite low since people don't know about the site yet. Over time you'd expect this figure to increase, but don't be surprised if it plateaus eventually due to stagnant content and a lack of Search Engine Optimization (SEO).

Time on Site
Another stat worth keeping an eye on is Time on Site.

This is the average amount of time people are spending looking around your site before they decide to go off and do something more important like watch Big Brother or floss their teeth. For most websites a low figure (e.g. 10-20 seconds), means something is seriously wrong. A low Time on Site tells us people just aren't hanging around for some reason.

There are some instances where a low Time on Site is not necessarily a bad thing. Imagine a website with weather information, it doesn't take that long to check if it's going to rain or not.

Bounce Rate
Bounce Rate appears on page 2 of the report and is of vital significance.

From a visitor's perspective, Bounce Rate means: 'I came, I looked around, I left.'

So, a high Bounce Rate means most visitors came and left straight away. In almost all circumstances this is a bad thing (especially if you are selling goods online). It indicates people simply aren't finding what they thought they would at your website.

If you can't convince people to stick around, what hope do you have (even if you have bucket-loads of traffic)? This is why some would argue that this is the single most important metric to monitor, even above unique visitors.

Unfortunately you can't make everyone stay at your website (unless you're willing to stand behind them with a cattle prod). Simply put, if you don't have what they want, they will leave. If you can maintain a Bounce Rate of around 40%, that's good.

There are rare occasions when a high Bounce Rate isn't as damning as you may think. An example would be news aggregate site where there are headlines linking off to other websites.

You can go much deeper into how Bounce Rate metrics work, but that's beyond the scope of an introductory guide such as this one.

Top Traffic Sources
Top Traffic Sources can be found on page 3 of the Analytics report.

This shows you where your traffic is coming from, or another way to look at it is how people are getting to your website.

You'll see something called Direct Traffic or (direct) ((none)), this refers to people who actually type www.acme.com into the browser address bar (i.e. they didn't find you through Google).

The google.com (referral) stat shows people coming from Google sources other than their search engine (e.g. Google Groups).

It's not unusual for a large portion of your traffic to originate from Google. People are searching for a particular term and Google is presenting your site as the best possible match.

The organic part in google (organic) means 'not paid for'. You got that traffic from Google on the merit of having good useful content people want. Traffic stemming from an AdWords click wouldn't be counted as organic.

The reason why traffic sources is worth paying attention to is because it can show how successful promotional efforts have been.

Take the PM4Web blog (which you are reading at the moment). I spent hours searching the Stack Overflow forum for posts on topics I have written about. If someone was asking for recommendations on what project scheduling method they should use, I would give a few suggestions plus refer them back to my full blog article.

The next week my Analytics data showed a distinct surge in traffic originating from the Stack Overflow domain. This meant that the time I invested in promoting my blog had paid off.

Traffic sources are also significant if you are spending money advertising on another website. If you are paying $200 a week to be listed on a directory site specific to your industry and no traffic is coming from that source, then you know to stop wasting your money and invested it in commemorative stamps instead.

Keywords
The Keywords segment is located on page 3 of the Analytics report.

This basically says: 'what words did people type into Google to cause it to send them to me.'

How is this information useful? If lots of people are coming to your website because of a particular topic, then it would make sense to encourage additional traffic by providing more of the same kind of content (i.e. give people more of what they like).

This may also affect your writing style, especially with regard to Keyword Density. Say you had a page on your website where you talk about roof repairs. On this page you use the word 'restoration' 16 times, and the word 'repair' 4 times. If your Analytics showed that most people found your site using the word 'repair' (rather than 'restoration'), that would be a good reason to go back and change things around. The theory being you should see an increase in future traffic since the page is now keyword rich for the term 'repair'.

Country/Territory
Page 4 of the Analytics report contains data about which country people are visiting from.

For some people this information is largely irrelevant. To illustrate this point; imagine you repair motorcycles, chances are your target market is going to be mostly local.

For others, this information can be very significant. Take this blog for example; I'm in Melbourne, Australia, but most of my readers are in the United States (yes, I am riding a Kangaroo as I type this article). To accommodate my primary audience I use American spelling instead of UK English (e.g. 'summarize' instead of 'summarise'). In addition, I adapt the language I use. I would say 'cell phone' rather than 'mobile' ('mobile' is our term for a cellular phone).

This data can be valuable for people selling goods online. If your current policy is to only ship within your own region, but you find most your traffic is coming from overseas, that could be a trigger to seriously reconsider your shipping practices.

Top Content
On the report's last page there's a segment called Top Content. These are the pages most people looking at.

Analyzing this data is important for a number of reasons. If you've just spent a great deal of time or money on a page, you can verify if it was a worthwhile endeavor by checking if people are visiting it.

The other major reasons is this: if people like something in particular, why not give them more?

This data can be helpful when deciding how to direct future upgrade efforts. If you had a very basic gallery page on your website and found that it was becoming increasingly popular, these statistics could be enough to justify upgrading it to something fancier.

The 'who Cares' Figures
There's some areas of the Analytics report which are hard to get excited about. For most people, knowing what browsers people are visiting with or their broadband speed is of little value.

There are situations where this can be important. Say if you found most of your visitors were on a really slow connection (cavemen or the Amish). This may prompt you to re-compress all the JPEGs on your gallery page, bringing the average photo size down from 60KB to 35 KB. That would be of little consequence to broadband users, but a significant improvement in user experience for dial-up visitors.

I'm being a little flippant implying that the Analytics report contains 'useless figures', we all know Google does an excellent job with the services they provide, all the figures in the report are going to be useful to someone.

The bottom line is to let the data educate and guide you. Making informed decisions based on metrics is always going to yield better results then guess-work alone.

Join RSS FeedSubscribe to RSS Feed.