Uh, it looks bad. Let's fix it. What do you mean "Justify It?"
September 28, 2011 5:33 PM   Subscribe

I work on a website and I need to justify new development for next year. Stuff like a site redesign and new features. I need to show the business case for doing it though, even though it looks like ass. But I have no idea how to do this.

I've been pretty shielded here from this stuff before and focused on execution, but there's no longer a person above me to do this stuff and now I'm It.

I can easily pull metrics from our site. And, I can make the determination that the problem is that people aren't staying on our site (lowered pageviews & high bounce rate) but how do I determine what the potential percentage increase in pageviews and the % decrease in bounce rates are so I justify making improvements?

I have two issues with this:

1. Am I supposed to just guess what the increase is? Or is there some research I can do to know if it's 1% or a 5% increase?

2. I don't know what excel formulas to use either. Math isn't my strong suit nor am I good at excel. Are there templates I can use?

I know I'm not the first person who had to do this, so how do I go about it? Is this even the right way to justify my work? Thanks.
posted by vivzan to Work & Money (5 answers total) 3 users marked this as a favorite
 
Look at your competitors. There are bound to be bigger ones in your industry with sharper looking sites. Keeping up with the Joneses, Inc. has traditionally been a driving force behind updating your company's online look.
posted by chrisfromthelc at 5:40 PM on September 28, 2011 [1 favorite]


As for your pre-question, does the business depend on pageviews and bounce rate for its bottom line?
posted by rhizome at 5:52 PM on September 28, 2011


Sounds like you're over thinking it a bit.

Your problem is:

You're seeing lower page views.

Your desired outcome

Increased Page Views, longer time spent on the site.

Your desired plan of action.

1. User interviews
2. A/B testing.
3. Increased Analytics. (something like Crazy Egg?)
4. Review Metrics
5. Implement Changes based on research


Business Case.
We have seen decreased traffic on our site leading to reduced Ad Sales/Revenue Sales (What type of site is this?) If we can increase our page views by 20% we should see a corresponding 20% growth in revenue. Site made X dollars last year we want to spend Y on increased traffic. Our goal is to increase revenues by Z. Z is less then Y so the improvements should pay for themselves.

Note if Z is greater then Y you have a harder sell.

Also don't say you'll increase traffic by some percentage. You haven't done the research to figure out what changes you can make and their impact. Just give a target goal. Also break the budget up into discrete steps with triggers for the additional budget. For example if your analysis can't prove a course of action that will see a significant increase in traffic, you don't get the money to make changes etc.
posted by bitdamaged at 5:53 PM on September 28, 2011 [2 favorites]


Response by poster: we make money via pageviews. the area of the site I'm responsible for is a big SEO traffic driver and if people aren't sticking around, it's a problem. Right now, we don't have a revenue problem. But, we're seeing usage of the product I'm responsible for trending downward. Some of it is related to people using our mobile apps instead, but some of it is that my area of the site is outdated compared to our competitors.

@bitdamaged thanks for the feedback.
posted by vivzan at 6:10 PM on September 28, 2011


Am I supposed to just guess what the increase is? Or is there some research I can do to know if it's 1% or a 5% increase?

This is pretty much the fundamental hole in using analytics for decision making. Knowing you have a problem is different than knowing how much you can change reality. It's even worse because for ad driven sites because all the standard sales driven metrics and tactics don't apply. If your information is highly relevant and informative, your bounce rate from organic search traffic will be huge!

So your task here is to find a number that is both possible and keeps you employed. I'm guessing you know how much you cost the company, so calculating out that end is easy. Whether 1 percent or 5 percent is more realistic is harder. One thing you can do is look at historical data: how much have things changed in the past, and how much of that is attributable to the firm's actions? That will inform your guesstimate.

If you really don't feel comfortable with using the past to measure tiny differences in a noisy signal, one other thing you might do is try a small change right now as a "feasibility study," similarly to bitdamaged's triggers. Brainstorm up twenty tweaks, choose the one you think has the best chance of working, and just put it up right now in an A/B test. If the testing says it harms or doesn't matter, try the next. If none of them work, well either you're hitting a wall, or you've found evidence for a more radical redesign.

One last ditch tactic: have you tried asking for last year's version of this analysis? It sounds like maybe the person who used to did this was fired, so maybe give it a careful look over if they even bothered with this in the past.
posted by pwnguin at 6:35 PM on September 29, 2011


« Older Is the $20 Las Vegas sandwich for real?   |   Hopefully it won't be a trail of tears. Newer »
This thread is closed to new comments.