img

Smart Bidding: Machine Learning Meets Advertising

Smart Bidding is a subset of automated bid strategies that use machine learning to optimize for conversions or conversion value in each auction—a feature known as “auction-time bidding”[1].

Google ads gives us several ways to bid for our ads depending upon our business and marketing preferences. At the elementary period of ads, bidding was so manual. Separate bids had to be involved for individual keywords or placements. With the introduction of mobile devices, we got to deal with enormous amount signals including device type, location, time, day, etc. All these signals, you see, have been too much for one human brain. Smart Bidding methods remove the guesswork and hard lifting out of bidding on Google Ads. To pick the best bid for each sale, it analyses billions of signals. You can thereby save time on manual activities and put more time on strategic efforts, while Smart Bidding helps your company reach its marketing objectives. [2]

Automated Bid Strategy

Bids for individual ad groups or keywords do not need to be updated manually. Google Advertisements calculates bids for our ads based on the probability that the ad will result in a click or conversion that will help you reach a specified business goal. Different types of automated bidding strategies can help us enhance clicks, visibility, and conversions. Smart bidding is just one of them.

Smart Bidding uses the advantages of goal-oriented automated bidding to conversions (clicks that result in sales or services) and conversion values. To capture the unique context of each search, it considers a variety of auction-time signals such as device, location, time of day, remarketing list, language, and Operating system. Smart Bidding employs a feature known as “auction-time bidding,” which establishes a bid for each query. [3]

Auction-time Bidding

Google utilizes the finest signals available to assist us place the best bid at the right time for the auction. This is referred to as auction-time bidding. It has the granularity needed to set a bid that is tailored to a user’s specific search query and context. [4]

4 Strategies of Smart Bidding

  1. Maximize Conversion Bidding: Maximize conversions determines bids autonomously to help you get the maximum conversions for your marketing while staying within your budget. Based on where, when, and how people search, bid adjustments allow us to show our advertising more or less frequently. Our prior bid adjustments are not used since Maximize conversions lets us improve our bids using real-time data. Only exception is that mobile bid adjustments can still be set to -100%. [5]
  2. Maximize Conversions Value Bidding: The total conversion value of our campaign is increased within our stated budget with the Maximize conversion value bidding strategy. Advanced machine learning is used in this approach to automatically fine tune and set bids. When we set up conversion metrics for our account, we specify the value we want to maximize, such as sales revenue or profit margins. [6]
  3. Target CPA Bidding: This is a Smart Bidding technique that fixes bids to maximize conversions (customer actions). We specify an average cost for each conversion when we build the Target CPA (target cost-per-action) bid strategy. Google Ads employs our Target CPA to set a bid based on the auction’s chance to convert when a customer searches for our product or service on Google. [7]
  4. Target ROAS Bidding: Every time a user searches for the products or services we’re advertising, this strategy evaluates and smartly estimates the worth of a prospective conversion. Practically, this implies that if the bid strategy indicates that a user search is likely to result in a high-value conversion, Target ROAS will place a high price on that search. This bid technique will bid low if it judges that the search is unlikely to result in a high-value conversion. [8] For example, let’s say we’re measuring sales for our online watch store and want to gain $5 worth of sales, for each $1 we spend on ads. The goal of $5 per sale is our conversion value, so we’d set a Target ROAS of 500%. This means for every $1 we spend on ads; we’d like to get 5 times that in revenue. To use Target ROAS bidding for search, we must have at least 15 conversions in the past 30 days. [2]

Guidance to Smart Bidding Selection

Maximize clicks is an automated bid strategy that sets our bids to help get as many clicks as possible within your budget. Target impressions share sets bids with the goal of showing our ad on the absolute top of the page. [10]

Why use Smart Bidding?

Smart Bidding is so effective to us with 4 key benefits.

  1. Advanced Machine Learning: Algorithms learn from large volumes of data to help us make more accurate predictions across our account about how different bid amounts affect conversions and conversion value. These algorithms consider a broader range of performance-related parameters than a single person or team could.
  2. Wide range of Contextual Signals: With auction-time bidding, we can optimize our bids based on a variety of indications. This contains features such as device and location.
  3. Flexible Performance Controls: Smart bidding enables us establish performance goals & personalize settings to our business goals like enhancing search bids to selected attribution model, setting device- specific performance goals for mobile, desktop and tablet with Target CPA.
  4. Transparent Performance Reporting: Reporting tools assist us gain a better understanding of our bidding performance and troubleshoot any problems promptly. These include bid strategy reports, status updates [11], campaign drafts & experiments, simulators, alerts & notifications, etc. [1] Apart from these smart bidding offers ability to choose conversion action at the campaign level, ability to adjust bids automatically during sales start or stop, ability to boost bids over multiple campaigns with a specific set of conversion actions, etc. [12]

Implementation

The process of implementing Smart Bidding can be discussed in 4 steps.

Step 1. Assessing the suitability of account 

We need to evaluate first whether our account is suitable for Smart Bidding strategy or not. There are some major reasons this strategy not fitting on google accounts.

  • Too new & inadequate data. Google’s advice is “30 conversions in the past 30 days” for Target CPA and “50 conversions in the past 30 days” for Target ROAS.
  • Conversions unsegmented. If the account has multiple sorts of conversions, it should be ensured that they’re all correctly categorized before starting a Smart Bidding plan.
  • Missing Data in Google Ads. We need to check whether any key pieces of data are absent which could misguide Google’s machine learning.
  • Unstable market. Some businesses operate in highly unpredictable markets, where a single day can have a year-long impact. Because of the desire for control, the volatility in some circumstances will dissuade you from ever utilizing Smart Bidding.
  • KPIs don’t align with smart goals. Currently Smart bidding covers website traffic (Maximize Clicks), low costs (Enhanced CPC), ROAS, and conversions.

Step 2. Selecting an appropriate Smart Bidding strategy

2 tiers of Smart Bidding:

Tier 2: Maximize Clicks, Enhanced CPC, and Target Impression Share Bidding

Tier 1: Maximize Conversations, Target ROAS and Target CPA

  • Google’s machine learning will examine your entire account and use the specific auction signals it has access to.
  • The importance of portfolio-level optimization cannot be overstated. After running some tests, it’s clear that it almost always outperforms keyword-level optimization.

Step 3. Establishing a Testing Strategy

  • Choose an approach that isolates success. For the finest tests, we expect around 80 and more conversions monthly. Less than this, it takes longer time to establish statistically significant results. Around 3 to 5 campaigns are recommended.
  • Testing on shopping campaigns. We could clone current shopping campaigns that have been manually designed, then split by location & executed at the same time.

Step 4. Optimizing Performance 

  • More data is better. The more data and the higher quality data provided; Google’s machine learning can work better. Portfolio bid strategy is recommended to be used wherever possible to make use of more data.
  • Changing targets when necessary. The downside with constantly modifying the target is risk of short-term performance declines as Google’s algorithm adjusts and learns.
  • Utilizing time gains to focus on other search activities. Previously bidding was time-consuming. Smart strategies have helped reduce the time taken. So, we can concentrate on other stuff like ad copy, keyword additions, strategic work, etc. [13]

Evaluating Smart Bidding Performance

The timeline to evaluate smart bidding is as follows:

The learning period is the time in which Google Ads gathers the performance data needed to optimize our bids. During this period, avoid major changes as much as possible to the campaign for example, budget change, landing page swaps, etc. It typically takes 1 to 2 weeks for a bid strategy to adjust. Post control period of 2-4 weeks is maintained. Conversion lag is there as most clicks don’t immediately result in conversions. Because customers may interact with our ads through multiple searches & devices.

Performance can be evaluated through any of the 3 methods: a) Drafts & Experiments b) Target metric comparison through the bid strategy report & c) Pre/post metric analysis. If bid strategy is tested by running an experiment, then the results can be found from the scorecard that populates at the top of experiment shown below.

If a bid strategy is directly added to the campaign, bid strategy report is the ideal evaluation method. This report enables us to realize how our strategies have performed for key metrics. [14]

Paid Search Ad Spend Growth

Spending in the US behind search ads are constantly on the rise since 2019 and it’ll go up in the next few years too. [15]

Conclusion

Smart bidding can do wonders in optimization of bidding strategies. More importantly it’s proved to be beneficial for large companies where the data is huge, almost regularly needs to be updated and the bidding strategy implementation is very much important. Because it saves big level of time consumption, workload of professionals, improves performance, provides transparent performance reporting & bidding precision, etc. But to gain the utmost values of it, we need to go step by step and utilize it accordingly. When done correctly, Smart Bidding is the smartest option for driving clicks, conversions, and ultimately revenue and profit.

  • https://support.google.com/google-ads/answer/7065882?hl=en&ref_topic=6294205%20%5B1%5D
  • https://www.youtube.com/watch?v=xNUSRL5kbYg%20%5B2%5D
  • https://support.google.com/google-ads/answer/2979071?hl=en&ref_topic=6294205%20%5B3-10%5D
  • https://www.adzooma.com/blog/smart-bidding-google-ads-guide/%20%5B11%5D
  • https://www.forbes.com/sites/forbesagencycouncil/2019/10/03/top-marketing-trends-for-2020/?sh=44c46b6113d5%20%5B12%5D
  • https://www.searchenginejournal.com/smart-bidding-guide/307705/#close%20%5B13%5D
  • https://www.youtube.com/watch?v=kriG3wT3xR8%20%5B14%5D
  • https://www.statista.com/statistics/739047/google-paid-search-ad-spend-growth-us-industry/%20%5B15%5D