The Pacing Algorithm Behind Facebook's Advertising System

From:Mr. Dong is in Silicon Valley.(Microsignal: donglaoshi-123)


Facebook last month reported its brightfourth quarter 2015 results, which included more than 1 billion daily users.Revenue reached $5.841 billion, while advertising accounted for $5,637 million, more frighteningIt's from the mobile advertising business that accounts for 80% of revenue.


The author on Facebook has also created a marketing page, in fact, to get a certain limit of internal friends of the advertising (Campaign) promotion costs, when surprised to run efficiently (too fast, less than an hour dozens of knives are gone), because they have done the advertising system, but also made a lot of comparisons, from the form of diversity, The integrity of the function, targeting users, real-time, data analysis report of the comprehensive indicators Facebook are industry leader, the following is their internal analysis of the advertising system Pacing algorithm, but also suitable for advertising technology entry to view.


Start with some online advertising terms


1. Each ad (Ad) has a bid (Bid) and has its actual click-through rate in a certain situation (Click-Through-Rate, CTR)

2、Ads are charged per click (Charge per Click, CPC), the strategy has one-price billing (First-Price, FP, i.e. how much an ad bids are billed for a single click) and second-price billing (Second-Price, SP, i.e. ads press a bid to pay the click price, more perp) All the time, GSP, Google, Yahoo, LinkedIn are all using this kind of)

3、Thousand show charge (Cost Per Mille, CPM, or RPM, R for Revenue), i.e. revenue in the event of a 1,000 impression slicing for a pay-per-click ad (one-price billing equivalent of 1000?CTR-Bid), or a thousand shows with a fixed price for display ads

4、Estimated click-through rate (predict CTR, pCTR) is the system's estimate of the likely probability of an ad's click before it is to be shown in a certain situation


Pacing is an algorithm in Facebook's advertising system that adjusts the pace of spending budgets, an analogy for runners: a premature sprint means no energy before the finish line, but a late sprint you might not have finished the race. Pacing guarantees that all advertisers will automatically allocate different advertising budgets under competitive conditions. Pacing is the core component of optimization that maximizes the return on investment (ROI) for advertisers.


How does The Facebook Pacing algorithm work?


We explain how Pacing works with an example. The following will involve clicking, the same idea can be applied to browsing, conversion, behavior, arrival rate, etc.


1、An advertiser wants to advertise a sports brand.With a budget of $10 per day and billed at CPC, we presuppose that each ad click will generate $5 in revenue.

2、When he/she creates an ad, do optimization for LINK_CLICKS (click on the link), bid_amount (bid) is set to $5, billing_event (trigger ingesting charge event) for LINK_CLICKS, based on these are real conditions. The target user base is male 25-35 years old.

3、Advertisers' profitises are the value generated by these clicks minus the budget spent.


For the sake of simplification, let's assume that the day's prices for these opportunities (clicks, presentations, etc.) are known in advance. This allows us to clarify more clearly the click price and revenue relationship. These assumptions are modeled based on the opportunities generated by the goal.


Here are three examples to understand the maximum value the Pacing algorithm provides to advertisers:


Scenario 1: When there's no Pacing algorithm


Without Pacing, advertisers' budgets are consumed clean lying out for a short time (potentially expensive clicks) at the beginning, intense competition in the early stages, no competition in the latter stages, resulting in a certain waste of resources. The blue dot in the figure below represents the opportunity for the ad to be shown, the yellow circle represents the ad to win the display, and the red line represents the price of the bid. The result is higher average costs, but advertisers get the most desired delivery based on ad settings. This is called accelerated delivery.

Total Ad Spends : $10
Total Click Value for Ads : 6 s $5 s $30
Value per click s $5
Budget : $10
Total Ad Revenue s $30 - $10 - $20


Case 2: Too low bid


In this case, the pursuit of the lowest click price, but advertisers' budgets are not used up in the end, and the final ad performance is the worst.

Total Ad Spends s $4
Total Click Value for Ads : 4 s $ 5 s $20
Value per click s $5
Budget : $10
Total Ad Revenue s $20 - $4 - $16


Case 3: Equilibrium State under the Pacing Algorithm


That's when advertisers get the most hits, get the most out of their money, and run out of daily budgets.

Total Ad Spends : $10
Total Click Value for Ads : $7 x $5 . . . $ 35
Value per click s $5
Budget : $10
Total Ad Revenue s $35 - $10 - $25


Simple formula

From the example above, when we used a steady bid (Pacing) throughout the period, the advertiser's value was maximized by not pacing or overbidding. To achieve its goals, Pacing expects to learn from other competitive ads with similar targets.Propose an optimized bid. This is a simple formula.

Final Bid (Per Show) - Optimised Bid (Per Show) - CTR where 优化竞价 <= max_bid

The decision to optimize the bid is at the heart of The Pacing algorithm, which includes a feedback system that allows Pacing to record.

CTR is the click-through rate. We use the same philosophy to look at the viewing rate (VTR) and conversion rate (CVR). The accuracy of these values increases steadily and is influenced by a variety of factors, such as ad type, audience, point in time, ad context, etc.

The Importance of Pacing

Pacing maximizes advertisers' profits on a given budget. It makes advertisers more real, and Vicky-Clarke-Groves (which is priced by calculating the sum of the losses that one advertiser would lose to other advertisers) has no benefit in deceiving value and doesn't have to think about how much the maximum bid should be set. This avoids special opportunities in the auction system.


Pacing guarantees a predictable delivery. This stable delivery stabilizes daily prices and helps advertisers gain fair access to the threshold for their target users.

FAQ

Q: My ad is not optimized correctly by the pacing algorithm, what might be the reason?


A: If the ad shows normal, there are two reasons, one is the most optimized price is too low, to ensure that your bid is within the proposed price range so that you have a chance to win a seat. For a competitive audience, you need to raise the price above the suggested price.


Another reason is that the target audience is too narrow.


If this ad is over-run, you may have a large audience but quickly consume your budget.


Q: How will Pacing be affected when I change my budget?

A: Pacing will calculate the new optimal bid, but this process will take a certain amount of time, during which time the bid is not optimal, so it is not recommended to frequently modify the bid and budget settings.


Summary: Pacing is about learning about the advertising competition environment within the same audience goal to determine the optimal bid.


From WeChat public number "Mr. Dong in Silicon Valley", focus on Silicon Valley trends, learn to grow together.



This article number 91, later want to read this article directly enter 91 can be.

Input m can be obtained to the article directory.

Today's WeChat public number recommends the
 


For more recommendations, please see15 technology-based public WeChat


Covers: Program Life, Algorithms and Data Structures, Hacking Technology and Network Security, Big Data Technology, Front End Development, Java, Python, Web Development, Android Development, iOS Development, C/C, .NET, Linux, Database, Operations, etc. Spread ingenuity computer learning experience, recommend computer excellent resources: click to go to15 tech-class WeChat public numbers of concern

Click to read the original and view the music festival song list