The Future of Google Ads

Written by Utkarsh Singh

In the dynamic world of digital marketing, keeping up with the evolving circumstances & trends are integral to success. The incorporation of Geo Location targeting in Google Ads campaigns is a step in this direction.

This innovative strategy has shown remarkable results, fundamentally transforming the way we manage our campaigns and boosting our traffic while reducing costs.

How do you then optimize the ‘when, where, and how’ for a new user to come into the system? Read on to know more…

Terminology

Traffic: Number of new customers coming into the system via Google Ads

Google Ads Campaigns: When you advertise with Google Ads, your ads can appear in different places across the web.

A campaign is setup using the following as input:

1. Objective: Main Agenda for your campaign. Google offers a variety of options here.

2. Conversion Goal: Focus area of the campaign. It refers to the metric you are targeting.

3. Campaign Type: The type of campaign depends on the type of business and geography.

4. Location: Target locations for the campaign.

5. Budget: Total spending depends on the number of users we want for the campaign.

Google Ads API: The Google Ads API is the programmatic interface to Google Ads, used to manage large or complex Google Ads accounts and campaigns

Last Mile: The last mile refers to the last stage in a process, in our case, it corresponds to hyperlocal delivery from a dark store to your home.

Bid/ Bid Adjustment: Bid is the cost given to campaigns at which they are supposed to bring new customers. Bid adjustment refers to the changing of bids depending on the change in goals. It is the target metric for any Google campaign.

Stress: Internal metric to identify the current status of a given dark store. High stress = Excess demand. Low stress = Excess capacity to serve orders.

Google Campaigns

Let’s face it, Google(and Meta FB) campaigns are one of the biggest parts of the cost component in an organization’s P&L. It ranges from 0.3% to 15% of the top line of business depending on the industry and stage of the startup. This lends an additional caveat to experimenting on such a high-priority line item.

While Google and Meta FB have cracked the nuances of the advertising ecosystem, there are corner cases that arise due to the hyperlocal nature of quick commerce business. While we are using it for Zepto, this methodology can be extended to any business which is spread over multiple geographies.

By using Google Ads API we are attempting to target the last 2 components of the campaign dynamically:

1. Location
2. Budget

Solution Architecture

We run 2 parallel campaigns for each city:

1. Store level geolocation optimization (location + radial distance in KMs).
   a. Since we have store-level data for stress we will use the same for campaign optimization.
   b. Adding and removing stores will help prevent unnecessary impact on nearby stores where stress levels are under control.
2. Use actual campaigns from the Google ads to maintain stress levels within optimum ranges. Stress is a real-time metric and is generated every 5 minutes. We will recalibrate the campaign using the latest stress value.
3. To achieve step 2 we will set Booster and BAU city-level campaigns for manipulating the traffic:
   a. Booster campaign — This will add the required Stores where we have lower stress and additional bandwidth to deliver orders. This will be a high CPA campaign designed to bring traffic to specific stores. It will drop the store as soon as we reach the stress threshold
   b. BAU campaign — This is designed to bring traffic to a specific city, now we will drop stores from this for a specific duration of time when the store is stressed.
4. For the setup of the experiment, stress will be calculated in real-time for each store:
   a. If store stress is between 80%-120% — BAU Campaigns will run with that store location in it, Booster campaign will not have this store’s geolocation added to it.
   b. If stress crosses 120% — Drop the store from the BAU campaign, the store is still not added to the booster campaign. This will put a stop to new users coming to this store for new orders.
   c. If stress goes below 80% — Add a store to the booster campaign while still being present in the BAU campaign. This will double the amount of new users coming to the specific stores.
5. All these actions are evaluated every 5 minutes to ensure that the system is working with maximum efficiency.

Performance campaigns (areas with stress)

Below is the graphical representation of the same phenomenon

Case 1: Where demand is reduced
Case 2: Where demand is boosted

Here’s how our new approach has led to positive outcomes:

Here’s how our new approach has led to positive outcomes:

1. Targeted Pincodes: We experimented with adding or removing dark store locations from our campaign and have a real-time impact on the traffic that comes from it.

• Excluding the stores with stress, we can immediately stop the traffic and we will get 0 campaign impressions on that store going forward thus helping reduce the stress levels and improve customer experience
• Adding a booster campaign for stores with Slack, we can immediately start generating incremental demand at that pincode increasing daily traffic

2. Streamlined Last Mile Traffic: By aligning our campaigns with real-time demand, we effectively streamlined the last mile traffic coming from Google campaigns and diverted it to stores that are slacked from the ones that are under stress, ensuring that our products and services reach the right customers at the right time. This positively impacted customer satisfaction, retention, and loyalty

3. Cost Savings: This approach helped us identify areas where ad spend was not yielding desired results. By eliminating or reducing investments in high-stress stores, we were able to save a considerable amount of budget, which can now be redirected to stores with the capacity to fulfill more demand or saved for the next campaigns

4. Dynamic Bid Adjustments: By analyzing data and demand patterns, we calculate stress in real time and adjust our bids dynamically. This means that our campaigns will receive maximum visibility during quieter times while ensuring optimal spending during peak traffic times.