Treat Your Bid Management Like Your Daily Fantasy Players
December 18, 2015
In my article earlier this month on why PPC Life allows you to wear many hats, I stated how I get to do things at work that I enjoy doing in my spare time. I used FanDuel as an example as to how some things I enjoy doing are similar to my everyday work life. Well – when coming up with my latest bid strategy ideas I came up with direct ideas from how player analyses’ are done in an attempt to maximize points on daily fantasy sports sites.
Layering statistics from different time periods and leveraging recent statistics as the most valuable data when coming up with projections is how I typically set up player analysis. So, why not treat your keywords the same way in an attempt to maximize revenue? Combine long-term and short-term performance in order to come up with projections for the next two weeks.
This bidding strategy can be used on any account successfully, but it is highly recommended for situations where there are high valued products and low sales volume on a weekly basis. This type of bidding will allow you to utilize revenue numbers on more keywords rather than trying to project revenue for the next two weeks with the last two weeks worth of data – which may have revenue data on just a select number of keywords if you have lower sales volume per week.
Setting Up The Strategy
The first thing you have to do is download keyword level data with all of the metrics you want to utilize in your bid strategies from each date range you want to use in your formula. Let’s use the past six months and the past 2 weeks in this example. Once you download each – the last two weeks and the last 6 months worth of data – you then average out the past 6 months worth of data into 2-week averages, by dividing the statistic by 13 in this case. You can see what the data has looked like on average in the past 6 months in 2-week intervals.
Once this is done you can utilize VLookUps in order to get these averages in the same sheet as the past 2 weeks.
Then, the statistics we want to utilize overall would include the 6-month averages, but with emphasis on the last 2 weeks worth of performance. In this case we’re just using 50% of our projection on the last 2 weeks worth of performance, and 50% of our projection on long-term performance – last 6-month average.
What About Ratios?
I know a lot of paid search experts teach users to typically utilize a 1:1 ratio when doing manual bid changes. Meaning if you do bids every week, go back one week. If you do bids bi-weekly, go back two weeks in terms of data that you’d pull. This method obviously does not follow this practice.
When users speak to doing bid changes with a 1:1 ratio, they are typically utilizing impression share, average position, ROAS and CPL metrics in order to determine a percentage they are willing to raise their bids.
The information we are gathering based off these weighted averages we pulled revolves around one thing: how much revenue are we likely to get per click. This is something that past bid changes do not impact as much, as it is not considering how much each click costs based on our bids. It is just telling us how much revenue we typically receive when that keyword is clicked. Since this is the case, we can go back deeper in time and use weighted averages without hesitation. From there we can utilize the JAF: (Revenue/Click)/ROAS Goal to come up with a suggested bid.
What About Keywords With No Revenue In The Past 6 Months?
This is where the last 6 months of data comes back into play. Typically when doing bi-weekly or weekly bid changes we can visualize performance from the past few weeks, and see that certain terms just have not received a lot of traffic to prove themselves. We can then decide on a bid change based on impression share or average position. However, in this case, we can go back 6 months if the term has no revenue assigned to it in order to see just how many clicks this keyword has been racking up over the weeks without ever converting. Below shows the complete formula of the suggested bid.
This formula is: IF(WeightedAvgRevenue>0,(WeightedAvgRevenue/Clicks)/ROASGoal,(AverageOrderValue/(WeightedAvgClicks*13)/ROAS Goal.
If there is revenue associated with the keyword use the regular (Revenue/Click)/ROAS Goal, but if there is no revenue associated with the keyword multiply the average amount of clicks by 13 to get the total amount of clicks in the past 6 months, then use the average order value to show IF we got one sale out of these clicks, how much revenue would we be getting per click. From there we divide by the ROAS Goal.
Finalizing The Bid Changes
Woah – did anyone else notice that this guy writing this article is telling us to bid $200.48 on a keyword in this article? Does he work for Google? Why would we risk spending that much on one click?
Obviously, this would be crazy to implement a bid over $200 and I would not suggest that to anyone in any industry (as far as I know with the industries I’ve worked in thus far). Where this ‘suggested bid’ occurs often in the formula is if a keyword has just a couple clicks with the high Average Order Value within the account, the suggested bid will be extremely high as it estimates how much revenue we’d get per click if that keyword brought in one sale the past 6 months.
This is where the next step in the process comes into play. Setting maximum and minimum bid changes. Here is where I utilize impression share from the last two weeks, along with past performance (willing to bid higher on terms that have brought revenue in before). In these columns sort by things that have revenue and impression share in the past 2 weeks and decide on how high and how low you are willing to change bids accordingly. Once these modifiers are in place the new bid formula will be:
This shows that if the suggested bid is above our max – we move forward with our max change. If it is lower than our minimum – we then move forward with our minimum change. If it lands between, that is when we utilize the suggested bid directly:
Using these formulas allows you to look deeper into the past without completely ignoring the recent past, and assuring leverage on recent keyword performance. It also allows you to see how terms are likely to perform in the coming weeks based on both recent past and long-term performance. When using the 1:1 ratio when setting the max and min modifiers in terms of impression share you are taking into consideration the last bid changes when making the new changes.
Browse By Category
The Advertising Principles That Guide Facebook's Decisions
Facebook highlights the advertising principles that guide their decision making in response to recent events.
Smart Ways to Get More Marketing Budget and Better Tools
In this new live webinar, Kristin Vick from Hanapin Marketing and Jeff Sauer from Jeffalytics discuss how marketers can ensure they have the budget they need to be effective with online advertising and get the right tools to make that argument.
A bi-weekly newsletter packed full of resources and strategies that will help make you a better PPC expert.
Hanapin Marketing | The PPC Agency of Experts Behind PPC Hero
The Hero Conf Difference
Of the 50+ search and social conferences you can choose from, we're here to show you why Hero Conf is a can't miss event for busy digital marketers like you.
Improve Your Facebook Ads With Creative Hub and Split Testing
With the updates to Creative Hub and the Split Test feature, one can now institute a basic ad creation and testing process directly within Facebook.
Work Habits & Resources for the Stressed-Out PPC Manager
Feeling a bit stressed this holiday season? Get tips and resources that will help you become more efficient and alleviate stress!