Tips and Tutorials

Stop Wasting Money—Why Waiting for BI is Expensive

How much does waiting cost you?

From a personal perspective, it might not be that hard to answer the question.  Add together time waiting in line, time waiting for someone else, and any other time you spend “waiting” time in a week, then multiply that number by your effective hourly rate.  

It might not be very scientific, but it’d be in the right ballpark.

But when it comes to waiting for information, it’s an entirely different calculation.  Sure, there’s some time spent physically waiting, but a lot of the cost is opportunity cost, not direct cost.

The cost of not doing something is often much larger than the cost of standing in line.

Capturing these opportunity costs certainly isn’t easy and may not even be possible, but there are some things you can do today to reduce the cost of waiting.  I’d like to share some of those with you.

Don’t wait for perfection

As Business Intelligence leaders who live and breathe numbers every day, it’s a constant focus to get the most precise, complete measurements possible.  

But sometimes, that’s exactly the wrong approach.

At TracFone, we have a data feed that we send Google regularly to inform our AdWords campaigns.  Google uses this info to, among other things, route traffic to our website based on many factors.  Providing this data is a major driver of traffic to our many customer-facing websites.

Less data = less traffic.

You’d think that getting the data as complete as possible would be the primary goal, wouldn’t you?  But in reality, that’s the wrong approach.

We’ve found that providing accurate, timely information to Google is a far better predictor of campaign success than completeness.

Don’t get me wrong.  Complete data would be wonderful.  But it would come at the expense of timeliness, and getting updated data to Google—even if it’s only 60% complete—is far more valuable than waiting two more days for more complete data.  

In this case, waiting would cost us literally millions of potential dollars.  Not to mention the money we’d be costing our future customers by not putting the best cell phone deal in front of them!

Measurement is important—but don’t make people wait for it

As BI practitioners, it’s ingrained in us to measure everything.

If it’s not being measured, it didn’t happen.

I’m not going to take a position at odds with that advice, but I do think it’s important to keep it in context.  Just measurement isn’t enough, of course.  You want to take the time to build the right analytics on top of your products.  

If it’s fast and easy to build those measurements, great.  You’re done.  But if it’s not—if it’s going to take engineering or database work, for example—don’t let it hold up the actual product.

Measurement is critical, but often not at the expense of waiting.

Instead, I recommend this three-phase approach:

  1. No Analytics.  Release with no analytics if necessary.  Often the usage patterns of a new product aren’t very valuable until the users settle into a long-term usage pattern, so making them wait for measurement analytics isn’t in anyone’s best interest.
  2. Quick Analytics.  As soon as feasible, release a basic set of measurements.  These may not be the measurements you’ll ultimately settle on, and are often consistent across products.  Number of hourly/daily users, average engagement time, and number of page views are all good, generic metrics that work for many types of products.
  3. Long-term Analytics.  As soon as feasible, start measuring what you really want to measure.  This should have little if any impact on end users.  They’ve been getting value from your product since release, after all...all you’re doing is measuring that value now.

Over many projects, I’ve found that this incremental approach to measurement can significantly reduce the wait time my customers experience.  And that time has enormous value to them.

Proactive improvement

It’s not always the customer of the data product that does the waiting, of course.  The trend over the past couple of decades toward shared-service analytics organizations has also bred a generation of analysts who are comfortable waiting for their customers to tell them what they need.

But there’s a cost to me waiting to hear how I can provide better analysis too.  Not just the cost of my time, but the cost of an inferior product being in production longer than necessary.  

After all, isn’t it the people with the data who should know what’s working, and what could be improved?  That’s often the business intelligence group—not the business user—so why should we wait for them to tell us what can be improved?

If you’re not using analytics to offer proactive improvements to your products, it’s time to start.  But you don’t have to do it alone.

Instead, partner with the business.  Ask them what their business goals are, not just want they want the report or dashboard to look like.  

View what you’ve built through that lens.  You’re already measuring it, after all, so take the next step and think about what those measurements tell you.  How could you make it better?

Then work with your business partner to make the change.

The hidden costs of waiting don’t need to be hidden

As you can see, we’ve focused a lot on reducing the hidden costs of waiting at TracFone.  But these costs aren’t really hidden, they’re looking us right in the face every day.

And to the extent that they are hidden, they’re usually only hidden from you.  If you talk to the right people—develop a partnership with your business users—they’ll be happy to share those “hidden” costs with you.

I hope I’ve been able to provide some ideas about how to take your BI program to the next level by reducing the cost of waiting.  I’d love to hear from you in the comments.  How do you reduce the cost of waiting in your company?

 

×