Ads Draft Campaigns let you test changes at scale without impacting live performance; you can create, preview, and run experiments to compare bidding, creatives, and targeting, then apply winning changes to your campaign. Use the official Campaign Drafts and Experiments | Google Ads Scripts for automation and to manage iterations efficiently, ensuring your optimizations are data-driven and controlled.

Key Takeaways:

Drafts let you propose campaign changes without altering the original campaign, creating a safe workspace for edits.
Create an experiment from a draft to split traffic between the original and the variant, enabling A/B testing of changes.
The original campaign continues running during an experiment, so performance comparison is isolated and concurrent.
After the experiment ends you can apply the draft changes to the original campaign, keep the experiment, or discard the changes based on results.
Test one variable at a time, set an appropriate duration and traffic split, and evaluate statistical significance before applying changes.

What are Google Ads Draft Campaigns?

You use Draft Campaigns as a sandbox inside Google Ads to propose and scope changes to an existing campaign without altering live settings; drafts let you prepare bid strategies, new ad copy, targeting adjustments, or budget shifts, then validate them with Experiments (A/B tests) – for example, running a 50/50 split for two weeks to compare CPA and conversion lift before applying changes to your main campaign.

Definition and Purpose

A draft is a copy of your campaign where you can edit bids, budgets, keywords, ad creatives, and targeting so you can prototype changes safely; after you build a draft you can launch an experiment to route a controlled percentage of traffic to the draft and measure outcomes, helping you quantify impacts like CTR, conversion rate, or cost per acquisition before you overwrite the live campaign.

Key Features

Drafts isolate proposed edits from the live campaign, integrate directly with Experiments for traffic splitting, and expose comparative metrics (CTR, CVR, CPA) over a chosen date range; you can test single changes-like a 20% CPC reduction-or bundled strategies and choose the winning variant to apply, minimizing guesswork and rollover risk.

Create and edit bids, budgets, keywords, ad copy, assets, and audience targeting without affecting live performance.
Launch Experiments that split traffic (for example, 50/50 or other ratios) so you can measure incremental lift over a set period, commonly 2-8 weeks.
Compare core metrics side-by-side in the Google Ads UI or export data to Sheets/BigQuery for deeper analysis.
Roll changes into the original campaign if the draft outperforms, or discard the draft to maintain the status quo.
Knowing how drafts separate edits lets you avoid accidental spend changes and keeps historical data intact for accurate comparisons.

You can use drafts to run surgical tests (change only max CPC on 10 keywords) or strategic overhauls (new audience segments plus different creative), and then measure statistically meaningful differences – for instance, a 15% uplift in conversion rate from a two-week experiment with 40% traffic allocation is actionable evidence you can safely apply.

Fine-grained testing: change a single parameter (e.g., reduce max CPC by 20%) to isolate its effect on CPA and ROAS.
Bundled scenario testing: implement combined changes (new ad copy + audience bid adjustments) to simulate full strategy shifts before committing budget.
Time-boxed experiments: schedule start/end dates and run tests over typical conversion windows (2-8 weeks) to capture seasonality and attribution delays.
Data export & analysis: integrate experiment results with analytics tools to calculate statistical significance and cohort-level lift.
Knowing the experiment’s traffic split and time window lets you design tests that reach sample sizes sufficient for reliable decision-making.

Benefits of Using Draft Campaigns

You get a safe workspace to prototype changes-ad copy, bids, audiences or landing pages-without affecting live performance, and you can convert a draft into a controlled experiment to measure impact. That reduces costly rollbacks and supports team collaboration, versioning and faster decision-making; many advertisers use drafts to test 3-10 variations before full rollout, which keeps spend efficient while improving metrics like CTR, conversion rate and CPA.

Testing Campaigns

You can create side-by-side variants for A/B tests-swap headlines, CTAs, landing pages, or bidding rules-and run them as experiments with defined traffic splits (commonly 1-50%). Use conversion rate, CPA and ROAS to judge winners; for example, a 10% CTR uplift from a headline change often coincides with measurable CPA improvements within 2-4 weeks when traffic and conversions are sufficient.

Optimization Opportunities

You’re able to trial strategic optimizations such as switching to a target-CPA bid strategy, narrowing keyword match types, or layering in an audience segment, then measure downstream effects on conversion volume and cost. Drafts let you estimate lift before changing the live campaign, so you can prioritize optimizations that move key metrics-like cutting CPA by 15% or boosting ROAS by 20%-based on data from the experiment.

To operationalize this, create a draft, implement one structural change (bids, keywords, ad assets or audience), convert it to an experiment, and allocate a sensible share of traffic. Run tests for at least 2-4 weeks or until you hit a reasonable sample (commonly 100+ conversions) to reach statistical confidence, then apply the winning draft to the main campaign and monitor for regression over the next month.

How to Create a Draft Campaign

To create a draft, open the campaign you want to modify in Google Ads (Campaigns > Drafts & experiments > Drafts), click “Create draft,” give it a descriptive name, then make edits to ads, bids, audiences, or landing pages without touching live results. You can save multiple drafts, preview estimated impact, and later apply a draft as an experiment or push changes live when metrics validate the update.

Step-by-Step Guide

Follow this concise workflow: select the base campaign, create a new draft, change one or two variables (ad copy, bid strategy, audience), save the draft, run it as an experiment for 14-28 days or until you reach statistical confidence, then apply winning changes to the original campaign.

Draft Creation Steps

Step 1	Select campaign → Campaigns > Drafts & experiments
Step 2	Click “Create draft” and name it (e.g., “Q4 CPAs – Bid+Copy”)
Step 3	Edit ads, bids, audiences, extensions or landing pages
Step 4	Save draft and preview estimated impact
Step 5	Run as experiment (allocate 10-50% traffic) for 14-28 days or 100+ conversions
Step 6	Analyze results, apply winning changes, or iterate

Best Practices

Keep tests focused: change one variable at a time, run 2-4 ad variations, and set experiments to collect at least 100 conversions or 14-28 days of data so results are meaningful; allocate 10-50% of traffic depending on risk tolerance and campaign volume to balance learning speed with live performance.

When analyzing, track CTR, conversion rate, CPA and ROAS by segment (device, location, time) and use Google Ads’ experiment reporting plus third-party analytics for attribution clarity. Label drafts, document hypotheses (e.g., “increase max CPC by 15% to improve impression share”), and keep an iteration log – many accounts see 5-15% lift after two disciplined iterations. If an experiment shows a small lift with low volume, scale gradually rather than flipping all budget at once.

Managing Draft Campaigns

You should treat drafts as a collaborative staging area: assign edits to teammates, use change history to track who modified bids or creative, and link drafts to experiments for validation. For example, create a draft adjusting search bids +20% on top-converting keywords, run that as a 50/50 experiment for 7-14 days, then compare CTR, CPC and conversion rate before applying changes to the live campaign.

Editing and Updating

When you edit a draft, focus on measurable adjustments: swap three headline variants, change max CPC by ±15%, or add a new audience segment of users who converted in the past 30 days. Use bulk edits for large lists, preview ad variations in the draft pane, and rely on change history to revert any unwanted updates before they hit live traffic.

Transitioning to Active Campaigns

You can either apply a draft directly or convert it into an experiment with a traffic split-commonly 10-50%-to validate impact. Monitor key metrics for at least 7 days (14 for low-volume campaigns): CTR, conversion rate and CPA, then decide whether to apply the draft, tweak it, or discard it based on statistical and business thresholds.

When ready to push a draft live, preview the exact changes, set an experiment end date if you’ve been testing, and use the apply button to commit edits; Google Ads will replace live settings with the draft’s configuration. If performance worsens-say CPA rises by 15%-stop the experiment, revert to the original settings, and iterate within a new draft to isolate which change caused the degradation.

Common Mistakes to Avoid

When working with drafts you can easily introduce errors that skew results: testing too many variables at once, running experiments during atypical traffic weeks, or failing to set clear success criteria. Use focused hypotheses-test one element at a time-and log expected KPIs before launching. For example, if you change both bids and ad copy, you won’t know which drove a 15% lift.

Overspending on Tests

Allocate a modest portion of your budget to experiments-typically 10-20% of the original campaign spend-to limit downside. If your campaign spends $10,000/month, cap the draft at $1,000-$2,000 and run the test 2-4 weeks. Apply bid caps and daily limits so a losing variant can’t consume the entire budget before you analyze results.

Ignoring Performance Metrics

Don’t rely on impressions alone; monitor CTR, conversion rate, CPA and ROAS to evaluate changes. A 30% CTR increase with no conversion lift usually signals poor landing page experience, not a winning creative. Track quality score and impression share too, since shifts there reveal auction or relevance issues that drafts can mask.

Dig deeper by setting minimum sample thresholds: aim for at least 1,000 clicks or 100+ conversions before calling a winner, and run tests across 2-4 weeks to smooth daily fluctuations. Segment results by device, location and hour to catch hidden losses-if mobile CPA jumps 40% while desktop improves, you may need device-specific bids rather than a full-sweep change.

Advanced Strategies for Draft Campaigns

When you build drafts with intent, you can validate multi-variable changes-bids, creatives, landing pages, and audiences-before applying them. Try a phased approach: run a 15% bid increase on your top 10% of keywords for 7-14 days or until 50 conversions, test three headline variations simultaneously, and hold other factors constant. Doing so helps you quantify CPA shifts, CTR uplifts, and conversion-rate changes without impacting live performance.

Phased bid ramping: increase bids incrementally (5-15%) on top keywords and monitor CPA weekly.
Isolated variable tests: change only one element per draft to attribute results clearly.
Audience layering: create 3 tiers (past converters, recent visitors, in-market) and apply bid modifiers.
Landing page drafts: route 20-30% of traffic to a variant for 14 days or 100 conversions.
Automated rules simulation: run draft changes with simulated rules before applying to live campaigns.

Quick Reference

Strategy	Implementation Tip
A/B Testing	Test one variable, run 1,000 clicks or 50 conversions, 7-14 days minimum.
Bid Ramping	Increase bids 5-15% in stages; measure CPA and impression share weekly.
Audience Layers	Create high/mid/low intent segments and apply +15-30% bids to high intent.
Landing Page Variants	Route 20-30% of traffic to variant; track conversion rate and load time.

A/B Testing

You should structure A/B tests inside drafts to isolate impact: change only headlines, CTAs, or landing pages per draft and run for 7-14 days or until you hit 1,000 clicks or 50 conversions. Use CTR, CVR, CPA, and bounce rate as your core metrics and prioritize statistical consistency-if one variant shows a 10-20% lift in CVR with stable CPA, it’s a strong candidate to apply.

Audience Segmentation

You can segment audiences in drafts by intent and recency-examples: converters (past 30 days), engaged visitors (30-90 days), and in-market audiences-then apply different bid modifiers or exclusions to each. Track conversion rate and ROAS per segment and adjust bids by 10-30% based on performance to maximize return on ad spend.

For more depth, combine Customer Match with similar audiences and run RLSA tests: bid +20-30% for past converters, exclude low-value cohorts, and compare 1% vs 5% lookalikes. In practice, set up three audience drafts, run them concurrently for two weeks, and evaluate by cost per conversion and incremental lift to decide which layers to roll into the live campaign.

Conclusion

Presently you can leverage Google Ads draft campaigns to test and refine strategies before applying changes live; by simulating edits, measuring forecasted performance, and running experiments, you maintain control over budgets and bids while minimizing disruption. Use drafts to document hypotheses, iterate on copy and targeting, and deploy validated improvements to optimize your campaigns confidently.

FAQ

Q: What are Google Ads draft campaigns?

A: Draft campaigns are editable copies of an existing campaign that let you build and save proposed changes without applying them to live traffic. They preserve the original campaign settings and metrics while allowing you to modify bids, budgets, keywords, targeting, ad copy, and other settings. Drafts themselves do not run ads or accrue cost; they are a staging area you can later apply directly or use to create an experiment that splits traffic to test the changes.

Q: How do I create and save a draft campaign?

A: From the Google Ads interface go to Campaigns, select the campaign you want to modify, then open Drafts & experiments > Create draft. Give the draft a descriptive name, make the proposed changes within the draft editor, and Save. You can revisit and edit the draft as many times as needed. When ready, you can Apply the draft to overwrite the original campaign or Launch an experiment to test the draft’s changes against the original with a traffic split.

Q: How do drafts differ from experiments and when should I use each?

A: Drafts are private proposals that do not change live traffic; experiments (formerly “campaign experiments”) run the draft’s settings on a portion of live traffic to compare performance against the original. Use a draft to prepare and refine changes; use an experiment to measure impact under real traffic with a controlled split and reporting. Experiments provide comparative metrics and statistical insights, while a direct Apply immediately updates the original campaign without split testing.

Q: What are best practices for using draft campaigns effectively?

A: Isolate variables-test one or a small set of changes (e.g., bids, targeting, ad copy) so results are interpretable. Set an appropriate experiment duration and traffic split (e.g., 10-50%) to gather sufficient data without overexposing risk. Keep conversion tracking consistent, use a clear naming convention for drafts and experiments, document hypotheses and KPIs, and run tests long enough to account for seasonality and conversion lag. If using smart bidding, be aware that experiments can affect learning; allow time for bidders to adapt.

Q: What limitations or troubleshooting issues should I be aware of with drafts?

A: Limitations include one active draft per campaign, inability to change campaign type within a draft, and certain features or settings that may not carry over between campaign types. If you can’t create, apply, or launch an experiment, check account permissions, ensure no conflicting edits or active experiments exist, and verify shared budgets or portfolio bid strategies aren’t blocking changes. Ad disapprovals and policy checks apply when you apply a draft or run an experiment. If performance is unexpected, confirm conversion windows, attribution settings, and that the traffic split was implemented as intended.