Maximize Efficiency With Dayparting in Small Accounts

By , Senior Account Manager at Hanapin Marketing

93 SHARES

Working in PPC is all about making the most of what you have.  Whether it’s maximizing clicks on a budget, meeting (and surpassing) your Cost per Conversion goals, or just massaging your keyword bids to find the optimal position for your ads, efficiency is what it’s all about.  This is doubly so for a small account, where every last cent matters.

Sure, the stakes may be lower when you’re only spending fifteen dollars a day, but that doesn’t make it any less important to perform at a high level.  That means you need to play to your strengths.  And on a limited budget, one of the best tools at your disposal is to make use of ad scheduling and bid modifiers.

As a PPC Manager or small business owner, it can be difficult to wrestle with the idea that there are audiences that you don’t want to interact with.  But when you don’t have a lot to spend, you can’t waste valuable clicks on people who are less likely to convert – information seekers, bargain hunters, comparison shoppers, and the 2:00 a.m. insomniac crowd.  You can weed out these groups with many different tactics, whether it’s negative keywords, qualifying ad copy, or in this case – day parting.

To start with, you’ll want to get friendly with the AdWords Dimensions Tab (helpfully explained by Amanda here).  For the purposes of this article, we’re primarily concerned with two metrics:  hour of day, and day of the week.  Now, it’s time for an impromptu case study in PPC 101!

Dayparting by Hour of Day in the Dimensions Tab

For the UK-based account above (hours listed in Eastern Standard Time, mind you), you can see that there’s a very significant trend that appears when you look at the data parsed out hour by hour.  The hours of 4:00 a.m. EST to 1:00 p.m. EST were our clear winners in terms of overall conversion volume.  Impressions, clicks, and conversions all fall off significantly after that point.  That means, that out of the $500 budget allotted to this campaign for the month, roughly $156 was spent on hours that provided either no conversions or did so at a higher CPA than our best-performing times.

There’s a similar story that develops when you take a look at data parsed out by day of the week:

Dayparting by Day of Week in the Dimensions Tab

In the above example (still in the same account, over the same time range), we see that both Saturdays and Sundays are the weakest days in terms of overall performance.  Thursdays are similarly anemic, but still at a better Conversion Rate and CPA than either of the two weekend days.  In terms of raw spend, that equals roughly $131 that could be better spent on other days.

So, looking at just these two top-level data charts, what can we deduce?  We now know that the audience for this particular account is most likely to convert at the hours of 4:00 a.m. to 1:00 p.m. EST, and that they do so at a more efficient rate on weekdays – with Wednesdays performing the best out of all days by far.  What can you do to best capitalize on this information?

In this case, your best option is to utilize the Ad Scheduling function in the Advanced settings of the Campaign settings interface to get out of those time periods and days.  In this particular account, you could decide to go with something like this:

Basic Ad Scheduling in Campaign Settings

Here, we’ve gone with a simple block schedule going from 4:00 a.m. EST to 10:00 p.m. EST running Monday through Friday.  This is the most basic level you can take with the information gathered from the Dimensions tab.  Now, we’re concentrating on our best-performing days and better allocating our limited funds toward hours that are more likely to convert.  But you can also take it a step forward with another feature:  the Bid Adjustment system.

Advanced Bid Adjustments in Campaign Settings Tab

Now we’ve gotten a little bit more advanced with what we can accomplish.  With this setup (which is what we eventually went with), we’ll be running at 110% of our normal bids from the hours of 4:00 a.m. to 2:00 p.m. EST, drawing back to 100% until 10 p.m., and staying completely out of the hours of 10:00 p.m. EST – 3:00 a.m. EST.  In this case, we’ve also left a day parting window open from 3:00 a.m. to 4:00 a.m., but with a 90% bid modifier to lower overall bids in that inefficient hour.

The best part about this system is the flexibility it gives you in implementing the data found from the Dimensions tab.  You can get out of entire portions of the day based on your findings, or use it as a way to control inefficiencies while still allowing conversion opportunities to occur – just at a lower bid price.

So, in this case, what was the overall impact of this kind of change?  I’m not promising you the same kinds of results, but have a look at this:

Results show an increase in conversion metrics across the board.

In the three weeks before and after this change was made, conversion metrics across the board skyrocketed, just by playing to our strengths.  So, if you’re on a budget (or even if you’re not) and you need to make the most of what you have:  give ad scheduling a try!

Get more weekly links with our Fast Five newsletter! Five Fast Links in Your Email Every Friday.

Also send me a daily RSS digest

ACO_endad_Q2Compass2014_-ppc-hero

Twitter Facebook LinkedIn Google+ Email Print More
  • dlenef

    Eric, this is a simple, easy-to-grasp introduction for those who either haven’t considered or were confused about dayparting. Just a couple of clarifying points:

    1. I disagree that volume (impressions, clicks) should be a criteria for scheduling or bid modification. Most physical retail stores shut down at night due to volume, but that’s because the cost of staying open is more than the revenue when there’s less traffic. But there is no such overhead with PPC, so you might as well keep the lights on because you might net some leads or sales. Unless…

    2. What does matter is conversion rate, cost/conversion, and other positive metrics that demonstrate value generation. As long as visitors display positive behavior after clicking ads, you’ll derive value.

    3. The concept of basing ad scheduling on data and metrics is golden. But your example lacks statistical confidence with double-digit click counts and just a handful of conversions for each time segment. So I would caution readers to use at least an order of magnitude higher samples to determine scheduled display and bid adjustments.

    • EricCouch

      Thanks for the post! These are all great points to take in to consideration if you’re looking at using these kinds of tactics in your own account. Regarding your point about statistical validity, the trend has been ongoing for months, but this post was just for the ease of demonstration.

      That being said, sometimes you just need to make your account move in the right direction!

    • Graham S

      Glad someone mentioned sample size….

  • Daniel Eriksson

    i understand that this is based out if direct conversions from adwords search, into site and checkout. But what about adwords search to site, then generic search to sale. This can only be tracked on the multichannel funnel in analytics, and can only be proven on gross (if you turn down certain hours, overall gross should be same or improve to prove that what you did was actually right).

  • http://www.facebook.com/profile.php?id=516054358 Rowan Springfield

    Hi Eric, I’ve been doing quite a bit of analysis of accounts by hour of day, and by day of week. However, I find that if I (time consumingly) split out data by downloading hour of day data for week days and then download hour of day data for weekends, the trends are quite different, which I guess is expected. Have you ever done this? And also, do you know how I could do this in a simpler manner, as I end up having to create 8 reports and pivot them just for a month’s worth of data!

    • EricCouch

      I have – you can also do the same thing via an Account report from the Dimensions tab. You can add multiple segments (in this case, hour of day AND day of the week) and run a pivot table in aggregate in the same fashion as you did with the custom Analytics report. It’s definitely helpful for a smaller account to get that level of specificity with your dayparting information.