Attribution Models in Google Ads: How NYC Small Businesses Choose the Right Model for Accurate Tracking

If you run Google Ads for a NYC small business, you have probably watched conversion numbers fluctuate from week to week and wondered which clicks really earned the sale. The answer often lives inside something most small business owners ignore: the attribution model. Attribution decides which ad in a customer’s journey gets credit for the conversion. Pick the wrong model and you will under-fund your most profitable keywords and over-spend on the ones that just close already-warm leads. This guide walks Manhattan, Brooklyn, and Queens business owners through what attribution models in Google Ads actually do, the differences between them, and how to choose the right one without needing a data science team.

What Are Attribution Models in Google Ads?

An attribution model is a rule that decides how Google assigns credit for a conversion across the multiple ads a user may have interacted with before buying. A new NYC customer rarely converts on the first click. They might Google your service in the morning, see a remarketing display ad at lunch, and finally click a YouTube ad in the evening that drives them to convert. The attribution model decides whether the morning click, the display click, the YouTube click, or all three get credit.

According to Google Ads Help, attribution applies to all ads inside your Google Ads account — search, display, shopping, YouTube, and discovery. Without an attribution model, all credit defaults to the last click, which hides the contribution of earlier touchpoints in the funnel.

Why Attribution Matters for NYC Small Businesses

NYC small businesses run on margin. Every dollar spent on the wrong keyword or audience is a dollar not spent on a winning one. Attribution affects three downstream decisions: bidding, budgeting, and creative.

If Google’s Smart Bidding system thinks last-click is all that matters, it will bid hardest on bottom-funnel terms — your brand name, “near me” searches, and high-intent service queries. That works until you realize those clicks would have happened anyway. Meanwhile, the upper-funnel keywords that actually introduce new customers get starved of budget because they look “less efficient” under last-click. Switching to a model that credits earlier touchpoints reveals which terms are doing the heavy lifting of customer acquisition.

For a Manhattan accounting firm targeting both “accountant near me” and “how to set up an LLC in NYC,” last-click attribution would over-credit the “near me” query. A more distributed model would reveal that the LLC question is what introduces the customer in the first place. Budget allocation changes accordingly.

The Main Attribution Model Types

Google Ads supports several attribution models. Each takes a different view of which clicks deserve credit.

Last Click

100% of the credit goes to the last ad clicked before conversion. This was the default model for years. It is simple, easy to explain, and biased toward bottom-funnel queries. Most small business accounts unintentionally use this model because it is the default in Google Analytics.

First Click

100% of the credit goes to the first ad clicked. This model is useful if you specifically want to optimize for new customer acquisition rather than closing existing leads. It is rarely the right choice as a primary model but useful as a comparison reference.

Linear

Credit is split evenly across every ad clicked in the path to conversion. A customer who clicked four ads would assign 25% credit to each. Linear is a simple but honest model that gives no single touchpoint outsized weight.

Time Decay

Credit weighting increases the closer the click is to the conversion. The last click gets the largest share; the first click gets the smallest. This model rewards bottom-of-funnel performance while still acknowledging earlier touches. It works well for businesses with short sales cycles, like restaurants or fast home services.

Position-Based

Position-based gives 40% credit to the first click, 40% to the last click, and splits the remaining 20% between any middle clicks. The model recognizes that the first and last touches are usually the most important. It is a popular middle ground.

Data-Driven

Data-driven attribution (DDA) uses Google’s machine learning to calculate the actual contribution of each click based on your historical conversion data. It is the model Google now recommends for most accounts. It is also the most powerful — and the trickiest to qualify for. We cover it in the next section.

Data-Driven Attribution Explained

Data-driven attribution analyzes the patterns in your account to predict the lift each click adds to the conversion. Instead of using a fixed rule, it learns from your actual data — which campaigns, keywords, and audiences tend to appear before conversions versus non-conversions.

Why It Outperforms Static Models

Fixed models like last-click and position-based use the same rule for every customer journey. Real customer behavior is messy. A first-time customer needs a different attribution pattern than a returning visitor. DDA adjusts the credit dynamically for each path, which usually leads to smarter bidding decisions when paired with Smart Bidding.

Eligibility Requirements

In the past, DDA required a minimum number of conversions over the prior 30 days to qualify. Google has since opened DDA to all accounts, but the quality of the model still depends on the volume of conversions you have. According to Google Ads documentation on data-driven attribution, accounts with more than 300 conversions and 3,000 ad interactions per 30 days will see the most reliable model output. Smaller accounts still get a usable model but should monitor results carefully.

Setting Up Conversion Tracking First

DDA only works if your conversion tracking is accurate. If you have not yet set up proper Google Ads conversion tracking, fix that first. No attribution model can compensate for missing or duplicate conversion events.

How to Choose the Right Model for Your Business

The right attribution model depends on your sales cycle, conversion volume, and how much you trust automated bidding.

Short Sales Cycle

If most NYC customers convert within a day or two of first contact — typical for restaurants, salons, plumbers, and locksmiths — last-click or time-decay are both reasonable. The funnel is short, so the most recent click really does carry most of the weight.

Long Sales Cycle

For services with longer research and decision cycles — accountants, lawyers, contractors, web design agencies — last-click hides too much of the funnel. Position-based or data-driven are stronger choices because they credit the early touches that introduce the customer.

Low Conversion Volume

If your account generates fewer than 30 conversions a month, data-driven attribution may produce unstable results. In that case, stick with position-based or time-decay until you have enough data to switch.

How to Switch Attribution Models in Google Ads

Changing the attribution model is a five-minute task inside Google Ads, but the consequences can take weeks to settle, so plan the change deliberately.

Step-by-Step Walkthrough

Open your Google Ads account, click Tools and Settings, then Measurement, then Attribution. Pick the conversion action you want to update and choose the new model from the dropdown. Save the change. The model will apply to all future conversions immediately and Google will retroactively re-attribute the past 90 days of conversion data.

Expect a Performance Wobble

Smart Bidding will adapt to the new attribution data over 1-3 weeks. During this transition, you may see CPCs rise on some terms and fall on others as Google re-calibrates. Do not panic and revert in the first week. Give it a full 21 days before judging results.

Document the Change

Add a note in your Google Ads change history and your internal records. If results shift dramatically, you need a clean date you can point to as the cause.

Common Attribution Mistakes to Avoid

Most account owners do not actively choose an attribution model — they accept whatever was the default when they set up the account. That alone causes most attribution mistakes.

Not Auditing the Default

If your account was set up before 2021, it probably defaults to last click. Newer accounts default to data-driven. Audit your current setting today so you know what you are actually using. A surprising number of NYC agencies report seeing accounts they have inherited still running on last-click without anyone realizing it.

Comparing Models Without a Control Period

If you switch from last-click to data-driven on the same day you launch a new campaign, you cannot separate the effects of either change. Make one variable change at a time and let it run for at least 14 days before stacking another change on top.

Ignoring Cross-Device Journeys

NYC users frequently jump between phone, work computer, and home laptop. If you have not enabled enhanced conversions and cross-device tracking, your attribution model is operating on incomplete data — no matter which model you pick.

Measuring the Impact of a Model Change

After a model change, watch a small set of metrics to know whether the switch was worth it.

Cost Per Acquisition by Campaign

Look at CPA at the campaign and keyword level before and after the change. The total may stay similar, but the redistribution across campaigns is where you find the insight. Upper-funnel campaigns that looked unprofitable under last-click often look better under data-driven, while brand campaigns look slightly less efficient because they no longer hoard all the credit.

Conversion Volume and Quality

Track not just the number of conversions but the quality. If your lead-to-customer rate stays the same or improves, the change is working. If conversion volume goes up but the leads are weaker, the new model may be over-crediting top-of-funnel terms. Adjust accordingly.

Spend Distribution

Look at how Smart Bidding is allocating budget after the change. A healthy DDA-driven account often shifts 10-20% of spend toward upper-funnel keywords within the first month. If you see far more dramatic shifts, double-check your conversion tracking before assuming the change is correct.