1 - Floor price optimization

About floor price optimization

The floor price is a minimum price you are willing to sell your inventory. Adjusting floor prices for advertisers can increase revenue by 20% to 80%.

Simply setting a very high floor price to get a high eCPM would ignore the role of the fill rate: a high floor price leads to a low fill rate. Thus, floor price optimization is about finding the ideal combination of floor price and win rate to get maximum revenue.

In terms of optimization, Smart AdServer recommends:

  • setting a low floor price for placements (formats) to get a satisfying fill rate
  • spotting good opportunities based on buyer data (advertisers) and increasing floor prices for advertisers accordingly; a good opportunity occurs when the price offered by the buyer (advertiser), is a lot greater than the price he pays at the end (in the second price auction system, the paid price is the second best bid price plus one cent).

This article will show you

  • how to spot these good opportunities
  • create according floor price delivery rules to increase revenue

Video tutorial - English version

Video tutorial - French version

Bid landscape report

The bid landscape report allows to get a vision of bidding activity. It tells you

  • how many bids are offered within ranges of 10 cents (e. g. 50,000 bids between EUR 1.40 and EUR 1.50 and 5,000 bids between EUR 3.60 and EUR 3.70)
  • the average paid CPM of all these offered CPMs
  • the best floor price to maximize your revenue

Report creation - step 1

Report creation - step 1
  • go to Reports > Creation of a report (1)
  • select RTB+ (2)
  • define the Report's name (3); recommended name: "Bid landscape"
  • select a recent Period (4)

About the period

The period should be as recent as possible. Start with a longer period (a few days) and gather data; then, decrease the period until the current day (since 00:00).
With a long period you may

  • study advertisers who no longer bid on your inventory
  • not take into account that advertisers have changed their bidding strategy. For instance, an advertiser may offer EUR 8 on average. You may set a floor price of EUR 7.50, but then stop making revenue. What happened is that the advertiser changed its bidding during the long period and now maybe bids at EUR 3.

Report creation - step 2

Report creation - step 2

In the Columns section, select the columns in the following order:

  1. Offered CPM Range 0.1 (the bid landscape will be based on these price brackets)
  2. RTB Advertiser
  3. DSP
  4. Buyer
  5. Format
  6. Auction type (to be able to focus on open auctions only)

Note: If you selected these columns in the wrong order, read Report creation - step 4 below.

Report creation - step 3

Report creation - step 3

Switch to the Metrics section. The following metrics are preselected:

  • RTB+ Impressions
  • Revenue
  • RTB+ Bids

Add the metric: Average clearing CPM; this is the eCPM (the actual paid price) which will be compared with the offered CPMs to spot good opportunities)

Report creation - step 4

Report creation - step 4

On the right pane:

  • check if the Selected columns are in the order stated in Report creation - step 2 (see above); rearrange the columns (drag and drop) if appropriate
  • rearrange the columns Selected metrics (drag and drop), to have them in the following order: RTB+ Impressions, RTB+ Bids, Revenue, Average Clearing CPM

Click Generate report; wait until the report is created and download the csv file

RTB+ report metrics and financial data

For a comprehensive list of all RTB+ metrics and financial data, click here.

Prepare Excel template

Prepare Excel template
  • download the bid landscape Excel template here. The file has 3 tabs: Bid landscape, Graph, Data. Our Excel template has all the tabs and calculations ready to help you analyze your programmatic bid landscape.
  • switch to the Data tab and replace the sample data by the data from the downloaded Bid landscape report; Important note: Only copy the data; do not change the French headers (first row of the Data tab)
  • go back to the Bid landscape tab and update the data source of the pivot table

About the win rate

A high win rate (impressions / RTB+ offers) means that an advertiser wins auctions very often. Thus, the advertiser's average offered CPM is a lot higher than the 
other offered CPMs.

So, if the win rate is high, this means the average offered CPM of these bids might be a lot higher 
than their average paid CPM (since their paid CPM is EUR 0.01 + one of the other offered CPMs).

An average offered CPM which is much higher than the average paid CPM means there is an opportunity to set a 
higher floor price.

Example 1

Identify relevant advertisers

Identify relevant advertisers

Sort the advertisers in a decreasing revenue order, as shown above.

Filter high win rates

Filter high win rates

Filter the results for advertisers who have a win rate > 40%

40% is an arbitrary number which often leads to interesting opportunities. Increasing this number leads to better but less opportunities.

Determine high revenue

Determine high revenue

Filter for each advertiser and switch to the graph to see the bidding situation.

For the advertiser in the example above, there are two relevant revenue brackets:

  • (1) One bracket is between EUR 0 and EUR 1
  • (2) A second bracket is between EUR EUR 4.50 and EUR 6

Analyze offered CPM range 1

Analyze offered CPM range 1

Switch to the Data tab.

In the first bracket (EUR 0 to EUR 1), most revenue is generated with the Offered CPM range of EUR 0.3 (between EUR 0.3 and EUR 0.4). For this CPM range, the paid CPM is EUR 0.3. Therefore, there is no big difference between the Offered CPM and the Paid CPM.

Increasing the floor price (e. g. EUR 0.6) would result in bids with a higher paid CPM ( > EUR 0.6 instead of EUR 0.3). However, this would cut the revenue generated for an Offered CPM of EUR 0.3, which is the biggest source of revenue.

To conclude, there is no optimization opportunity for the first bracket.

Analyze offered CPM range 2

Analyze offered CPM range 2

In the second bracket (offered CPM range between EUR 4.50 and EUR 6, column 1), the Paid CPM (column 5), is consistently above EUR 3 (as opposed to only EUR 0.3 in the first bracket).

From these paid CPM differences (EUR 0.3 and EUR 3), we can conclude that we are probably dealing with 2 different floor prices for 2 different formats.

When seeing different floor prices for the same advertiser, you should add the format dimension to the pivot table to see which formats are involved. If the optimization does not target the good format, you may see negative impacts on the other formats.

Calculate floor price

Calculate floor price

The difference between Offered CPM range (column 1) and paid CPM (column 4) is biggest in the zone between EUR 5 and EUR 6 (Offered CPM, column 1).

Thus, setting a new bid floor of EUR 5 could be good optimization opportunity. To calculate the potential revenue, we apply an eCPM of 5 EUR to the sum of impressions in the zone between EUR 5 and EUR 6 (see red area 3 above). The sum of impressions is 74 542 impressions.

  • Revenue = Impressions * eCPM / 1000
  • Revenue = 74 542 * 5 EUR / 1000
  • Revenue = EUR 372,71

The revenue with the new floor will be at least equal to EUR 372 while the current revenue is EUR 251 (blue area 5), which is a revenue increase of 48% for this advertiser.

Use the filter to see which format or site affected to make the optimizations for the appropriate placement.

Note the new floor for this advertiser and proceed with the other relevant advertisers.

Example 2

Determine high revenue

Determine high revenue

In this example, most revenue is generated at an offered CPM of EUR 4.

Analyze offered CPM range

Analyze offered CPM range

The paid CPM is slightly above EUR 3 for all Offered CPM Ranges. Thus, the current floor price is probably EUR 3.

Calculate floor price

Calculate floor price

Increasing the floor price to EUR 4 is a reasonable optimization step.

To calculate the potential revenue, we apply an eCPM of 4 EUR to the sum of impressions in the zone between EUR 4 and EUR 6 (see red area above). The sum of impressions is 13 613 impressions.

  • Revenue = Impressions * eCPM / 1000
  • Revenue = 13 613 * 4 EUR / 1000
  • Revenue = EUR 54.45

Thus, increasing the floor price to EUR 4 would potentially generate a revenue of at least EUR 54 instead of EUR 45, which is a revenue increase of 21% for this advertiser.

Delivery rules

Delivery rules

Go to RTB+ > Delivery rules

Create delivery rule

Create delivery rule
  • define a Name - read below about naming
  • select the Effect Floor price and set the new floor price
  • select the advertiser
  • if appropriate, select placement (format, website, page)

Manage floor price delivery rules

For each target (advertiser), create a separate floor price rule. Even if the floor price is the same, create a separate rule for each advertiser. This way, you can clearly determine which floor prices work well and adjust the rule later.

Naming delivery rules with a single target

Naming pattern

[rule type] - [target category] - [target name]
  • [rule type]: corresponds to the effect of the rule (Block, Floor price or Transparency)
  • [target category]: advertiser, DSP etc.
  • [target name]: an advertiser, a format etc.

Examples

  • A floor price of EUR 5 for the advertiser "Nike.com" => "Fp - Adv - Nike.com"
  • A floor price for the 300x250 format => "Fp - Format - 300x250"
  • A block rule for the DSP "Google DBM" => "Block - Dsp - Google DBM"

Naming delivery rules with multiple targets

Naming pattern

[rule type] - [target category]
  • [rule type]: corresponds to the effect of the rule: Block, Floor price or Transparency
  • [target category]: advertiser, DSP etc.

In delivery rules with multiple targets, do not write the names of all targets into the rule's name.

Example

  • A block rule for multiple buyers => Blocklist - Buyers
  • A block rule for multiple DSPs => Blocklist - DSPs

Delivery rule impact report

Delivery rule impact report

You should check the impact (revenue increase or revenue loss) of the delivery rules on a regular basis.

Reach out to your Yield manager or Technical Account Manager to receive the delivery rule impact report on a daily basis (please provide your network name and the recipient's e-mail address).

With floor price adjustments suggested by this report, you can eliminate most negative effects of delivery rules.

The delivery rule impact report lists:

  • each delivery rule
  • its rule effect (block or floor price)
  • the day
  • the effect of the rule (red area above); the effect can be positive (revenue increase) or negative (revenue loss)

You can create a pivot table with this data for further analysis. 

If you detect a delivery rule causing a negative result (revenue loss), you should delete the delivery rule in the user interface.

optimisation

Was this article helpful?
0 out of 0 found this helpful
Powered by Zendesk