The short answer
For a dental group of about 3 to 50 or more offices, name the pilot offices and the comparison offices before anything changes. Write the success read first, using the columns on the per-office marketing report. Give each pilot office its own office slug, its own tracking, and a campaign name that matches that office. Run the pilot for a length the group chooses and writes down. Keep Google Ads, GA4, Search Console, the Business Profiles, and the domain with the group from the first day, and give the firm user access. At the end, decide to roll out, extend, or stop from that written read. None of the Google pages below publish a pilot length in weeks, a lead count, or a percent that means the trial worked.
This guide is for a marketing lead who wants to try a firm, or a new in-house approach, at a few doors before the rest of the group follows. It covers the list of offices, the read you write first, and the decision at the end. A wider scorecard, without a league table, is in how to choose a DSO dental marketing company. The map of the other group jobs is marketing a multi-location practice or DSO.
Pilot offices and comparison offices
Pick both lists before the first edit. Similar offices make the later read easier to talk about. An office that is moving, changing its public name, being acquired, or changing providers is a poor pilot row, because those events already change the profile, the page, and the phone. Use the worksheets for those events instead of forcing them into the trial: the post-acquisition transition, rebrand, co-brand, or keep the local name, and provider pages and practitioner profiles.
Pilot office
An office the group names before any change, where the agency or the new in-house approach will run. Office A is a label on the sheet. It is not a patient and not a clinician.
Comparison office
An office the group also names before the change, which keeps the prior approach. The written read sets the pilot rows next to these rows. Read it as a side-by-side view of these offices only.
Leave it out
An office in the middle of a move, a rebrand, an acquisition, or a provider change. Those shifts belong on their own worksheets. They are not a clean pilot row.
Exception line
Hours, phones, or staffing that changed at a named office during the dates. Write the fact on the exception line. Do not fold it into the result.
| Label | What the group writes down |
|---|---|
| Pilot office | An office the group names before any change, where the agency or the new in-house approach will run. Office A is a label on the sheet. It is not a patient and not a clinician. |
| Comparison office | An office the group also names before the change, which keeps the prior approach. The written read sets the pilot rows next to these rows. Read it as a side-by-side view of these offices only. |
| Leave it out | An office in the middle of a move, a rebrand, an acquisition, or a provider change. Those shifts belong on their own worksheets. They are not a clean pilot row. |
| Exception line | Hours, phones, or staffing that changed at a named office during the dates. Write the fact on the exception line. Do not fold it into the result. |
Three offices and fifty offices use the same columns. A larger group has more rows. It does not get a private list that only the firm can see. If hours, the phone, or staffing change at a named door during the dates, that fact is an exception line. It is not evidence that the approach worked or failed.
What the pilot measures
The success read is the per-office report you already defined: calls, qualified calls, form leads, online bookings, Business Profile actions, ad cost for a campaign named as one office, and new patient starts from the practice's own records. Same definitions. Same start. Same end. If a cell was empty in that report because the practice did not export it, it stays empty here too.
Per-office report
The same columns, the same start date, and the same end date for every named office. The success read is those columns, written down before the pilot starts.
Baseline
The same columns for a period that ends before the change. The group picks that period. Export it before anyone edits a campaign, a page, or a profile.
GA4 comparison
A comparison puts subsets of one report next to each other, using dimension conditions. Values inside one condition use OR. Separate conditions use AND. Partial match is case sensitive. Use the office slug spelling you already publish.
Search Console
The Performance report defaults to the past three months. Change the date filter so the pilot and the baseline use the dates on the sheet. Read clicks, impressions, click-through rate, and average position for the pages you named.
Business Profile
An owner or a manager can set a date range on a verified profile. The view includes organic results and Google Ads. Call-button clicks, website clicks, and direction requests stay in the profile column. The searches figure is updated at the start of each month and can take up to 5 days to appear.
Phone calls from ads
A call from an ad, a call asset, or a location asset can be a conversion when it lasts at least the minimum length you set. A click on a phone number on a mobile site is a click, not the call. Keep that rule the same from the baseline through the pilot.
| Read | What to keep the same |
|---|---|
| Per-office report | The same columns, the same start date, and the same end date for every named office. The success read is those columns, written down before the pilot starts. |
| Baseline | The same columns for a period that ends before the change. The group picks that period. Export it before anyone edits a campaign, a page, or a profile. |
| GA4 comparison | A comparison puts subsets of one report next to each other, using dimension conditions. Values inside one condition use OR. Separate conditions use AND. Partial match is case sensitive. Use the office slug spelling you already publish. |
| Search Console | The Performance report defaults to the past three months. Change the date filter so the pilot and the baseline use the dates on the sheet. Read clicks, impressions, click-through rate, and average position for the pages you named. |
| Business Profile | An owner or a manager can set a date range on a verified profile. The view includes organic results and Google Ads. Call-button clicks, website clicks, and direction requests stay in the profile column. The searches figure is updated at the start of each month and can take up to 5 days to appear. |
| Phone calls from ads | A call from an ad, a call asset, or a location asset can be a conversion when it lasts at least the minimum length you set. A click on a phone number on a mobile site is a click, not the call. Keep that rule the same from the baseline through the pilot. |
Analytics comparisons evaluate subsets side by side. Build the condition on a dimension you already collect, such as the session campaign or the office parameter, and keep the spelling identical, because a partial match is case sensitive. That article is about subsets. It points onward to comparing date ranges and does not, on this page, set a dental baseline. Export the baseline and the pilot as two reads with the dates you wrote down. Do not wait for a saved comparison to invent an office that was never tagged.
URL builders say to set utm_source, utm_medium, and utm_campaign together, and that the values are case sensitive. When the campaign is one office, the campaign value is that office slug. utm_content is the parameter Google describes for telling creatives apart. Use it for the slug only when one campaign still covers more than one door, and keep that cost on the exception list until the campaign is one office. Parameter values are not a place for a patient name or a chart number.
Search Console's Performance report shows clicks, impressions, click-through rate, and average position. The default view is the past three months. Use the date filter so every office page you are judging uses the pilot dates, and use it again for the baseline dates. The report can group by page. A query list is not a substitute for the office row. The article points to a further page for advanced filtering and comparison. This guide cites the overview you can open directly.
Business Profile performance is for a verified profile, read by an owner or a manager. You can set a date range. Only metrics that apply to that business are shown, and the figures include views, searches, and actions from organic results and from ads. Calls in that view are clicks on the call button. The searches metric updates at the start of each month and might take up to 5 days to show, so an empty searches cell inside that window is not a finding. Phone call conversion tracking is a different count: a call from the ad can be a conversion when it meets the length you set, and a tap on a number on a mobile site is only a click.
Paid search inside a pilot
An experiment inside one office's campaign is a cleaner read than a gap between two offices. Office A and Office B already differ in hours, providers, and demand. Google's Experiments page says to test one variable, to write the hypothesis against a business goal, and to pick one or two metrics before the test starts. A pilot result describes the pilot offices, so test again before you roll it out to the rest.
A custom experiment, on Google's setup page, is built from an original campaign and shares that campaign's traffic and budget. The page describes custom experiments for Search, Display, Video, and Hotel campaigns, not for App or Shopping campaigns. You can schedule up to 5 experiments, and only one runs at a time. A note on the campaign picker says the campaigns shown are active, use Smart Bidding, include exact or phrase match keywords, are not already in another experiment, and are not running on the Display Network. Read that note against the live campaign before you treat a Display or Video campaign as eligible.
For Search campaigns, the split can be cookie-based or search-based. In Display campaigns, Google says it always uses a cookie split so a person sees either the experiment or the original. If you use a cookie-based split and an audience list, the same page says the list should have at least 10,000 users, and that fewer users may make the result less accurate. That figure is about the audience list on the experiment. It is not a count of patients an office should have, and it is not a sample size for the group.
The setup note says the experiment arm splits the budget evenly between the control and the treatment. The same page also says you can choose what share of the original campaign's budget the experiment uses, and that a campaign which already has a shared budget cannot be the base until that shared budget is removed. A change on the original campaign is not copied onto the experiment. Editing either one while the experiment runs can make the result harder to interpret. The Experiments page adds a related caution: avoid changes to the base campaign unless sync is on.
Smart Bidding is auction-time bidding for conversions or conversion value. The strategies named there are Target CPA, Target ROAS, Maximize conversions, and Maximize conversion value. Google says some of them work from historical conversion data, and that a fair read uses a longer period with at least 30 conversions, such as a month or longer, and 50 conversions for Target ROAS. That is Google's note about reading Smart Bidding. It is not the length of this pilot. The group still writes its own length.
Bid strategy status can show Learning after a new or reactivated strategy, after campaigns, ad groups, or keywords are added or removed, or after the conversion goal changes. Google says Learning never lasts more than 7 days. The same page says changes to budgets, targets, creatives, or targeting such as location or audience do not remove historical data and do not trigger Learning. If the status is Learning, record it. Do not treat those 7 days as the pilot, and do not treat a budget edit as if it had started that status.
Average daily budget is the amount you set for a campaign per day. For most campaigns, Google says spend on a given day will not pass two times that amount, and spend in a given month will not pass 30.4 times that amount. The experiment spends inside the original campaign's budget. It does not create a second budget number for the office. Shared budgets, and which campaigns can share one, are in DSO Google Ads budgets and call attribution.
Location assets can show the address, a map or distance, and a call button for a store the advertiser owns, using the Business Profile. If a location is marked closed, those assets do not show. Use the asset for the pilot office's own profile. Call assets add a phone number to the ad. The more specific asset is the one that serves: an ad group asset overrides a campaign asset, which overrides an account asset. When a location asset and a call asset are both eligible, an ad that features a specific location may ring that location's number instead of the call-asset number. Write down which number the pilot office is using. Do not leave that choice inside the firm's account only.
SEO and Business Profile work
Google's page on whether you need an SEO says an audit should give realistic estimates of the improvement and of the work. It says no one can guarantee a number one ranking. It also says that, at the audit stage, you grant read access to Search Console and not write access. The page asks a firm what results it expects and on what timeframe. It does not state a timeframe.A pilot that includes SEO work still ends on the written read, not on a ranking promise.
Profile edits during the same dates are part of the record. A new primary category, a new phone, or a name change is an exception line next to that office slug. The performance date range should match the sheet. Because the performance view mixes organic results and ads, do not add those actions to the ad-cost column or to the call-conversion column. If the firm wants write access to Search Console for the pilot, that is a separate grant from the read-only audit Google describes, and the group remains an owner.
Ownership and access
On the first day, the group owns the Google Ads account, the GA4 account, Search Console, each Business Profile, and the domain. The firm receives user access it can lose without taking the asset with it. The role names, the primary owner, and the order for removing a user are already in switching a dental marketing agency and profile governance.It requires that they are true before the pilot starts, not after a renewal conversation.
Write the exit before the first change: which exports the group keeps, which users come off, and who remains on the domain. Ask your own counsel to read that exit before anyone signs.
Roll out, extend, or stop
The group writes its own bar. This table says what has to be true before each decision. It does not set a percent, a lead count, or a payback period.
Roll out
The written read is filled for the dates you set. Exceptions are listed, not hidden. The group still owns the accounts and the profiles. Leadership chooses to try the same approach at other offices. Each new office gets its own row and its own read.
Extend
A date, a tag, or an exception is still open, or the group wants another period and writes that period down before it starts. Ownership stays with the group. Extending is not a new success rule.
Stop
The written read does not support more offices, or the group cannot open the accounts, the profiles, or the sheet without the firm. Stopping is a decision on the sheet. It is not a claim that the work had no activity.
| Decision | What has to be true |
|---|---|
| Roll out | The written read is filled for the dates you set. Exceptions are listed, not hidden. The group still owns the accounts and the profiles. Leadership chooses to try the same approach at other offices. Each new office gets its own row and its own read. |
| Extend | A date, a tag, or an exception is still open, or the group wants another period and writes that period down before it starts. Ownership stays with the group. Extending is not a new success rule. |
| Stop | The written read does not support more offices, or the group cannot open the accounts, the profiles, or the sheet without the firm. Stopping is a decision on the sheet. It is not a claim that the work had no activity. |
Applying a custom experiment to the original campaign, or ending it, is a control inside that campaign. Google's setup page says applying the experiment keeps the performance data on the original, and that you can also end the experiment. That button is not the group's rollout decision. Rollout is the separate choice to use the approach at offices that were not in the test.
Red flags
These are checks on a proposal or on the read a firm hands you. They do not name or rank any firm.
-
1. The firm picks the pilot offices after it has seen the numbers.
The offices are named first, then the baseline is exported, then the change starts. Choosing the doors after the data is in is a different exercise from a pilot.
-
2. There is no baseline.
A pilot period with nothing to set it next to is a slice of activity. The sheet needs the same columns for a period that ends before the change.
-
3. The metrics change while the pilot is running.
The success read is the column list you wrote down. Swapping qualified calls, adding a new cost column, or dropping an office mid-way means the read you planned is not the read you were handed.
-
4. The only copy of the result is inside a vendor dashboard.
The group should be able to open the per-office sheet, the Ads account, Analytics, Search Console, and each profile. A slide that cannot be tied to those screens is not the read.
-
5. The accounts were created under the agency.
The group owns Google Ads, GA4, Search Console, the Business Profiles, and the domain from the first day. The firm gets user access. An account that exists only on the firm's login is an ownership problem, not a pilot result.
-
6. Someone promises a ranking or a guaranteed lead count.
Google's hiring guidance says no one can guarantee a number one ranking, and that an audit should give realistic estimates of the work.
Questions for an agency
Put the same questions in the RFP. A useful answer names a sheet, an account, or a person at the group. A weak answer describes a pilot without saying which offices or which file. These are procurement questions. They are not legal advice and not a template.
1. Which offices are the pilot offices, and which are the comparison offices?
A useful answer names
A list written before the work starts, with an office slug on each row, and a note if any door is mid-move, mid-rebrand, mid-acquisition, or changing providers.
Ask a follow-up if you hear
"We will pick the offices once we see the data," or a single office chosen because it already looks strong.
2. What will the success read contain, and what dates will it use?
A useful answer names
The per-office report columns, one start date, one end date, and a baseline export from before the change. The group can name the file.
Ask a follow-up if you hear
A new dashboard metric, a percent that was not in the column list, or dates that differ by office.
3. How long will the pilot run?
A useful answer names
A length the group wrote down. Google's Learning status never lasts more than 7 days, and that status is not this length. Smart Bidding's own note about conversion volume is not this length either.
Ask a follow-up if you hear
A week count presented as a standard for dental groups, or a promise that the length alone produces a result.
4. If paid search is in the pilot, where does the test sit?
A useful answer names
A custom experiment on that office's own campaign, or a written reason the campaign cannot host one. The experiment name, the split, and the base campaign are visible in the group's Ads account.
Ask a follow-up if you hear
Two unrelated offices compared as if they were one experiment, or a test that lives only in a vendor tool.
5. Who owns the accounts on the first day, and what is the exit?
A useful answer names
The group owns Ads, GA4, Search Console, the profiles, and the domain. The firm is a user. The exit, including what is exported and which users come off, is written before the pilot starts.
Ask a follow-up if you hear
"We will transfer ownership if you stay," or logins that only the firm can reset.
6. What will you do with a ranking claim or a lead guarantee?
A useful answer names
Neither is a deliverable. The firm can show the sheet, the date filter, and the campaign. Google's hiring page says no one can guarantee a number one ranking.
Ask a follow-up if you hear
A number one ranking, a guaranteed lead count, or a special relationship with Google.
| Question | A useful answer names | Ask a follow-up if you hear |
|---|---|---|
| 1. Which offices are the pilot offices, and which are the comparison offices? | A list written before the work starts, with an office slug on each row, and a note if any door is mid-move, mid-rebrand, mid-acquisition, or changing providers. | "We will pick the offices once we see the data," or a single office chosen because it already looks strong. |
| 2. What will the success read contain, and what dates will it use? | The per-office report columns, one start date, one end date, and a baseline export from before the change. The group can name the file. | A new dashboard metric, a percent that was not in the column list, or dates that differ by office. |
| 3. How long will the pilot run? | A length the group wrote down. Google's Learning status never lasts more than 7 days, and that status is not this length. Smart Bidding's own note about conversion volume is not this length either. | A week count presented as a standard for dental groups, or a promise that the length alone produces a result. |
| 4. If paid search is in the pilot, where does the test sit? | A custom experiment on that office's own campaign, or a written reason the campaign cannot host one. The experiment name, the split, and the base campaign are visible in the group's Ads account. | Two unrelated offices compared as if they were one experiment, or a test that lives only in a vendor tool. |
| 5. Who owns the accounts on the first day, and what is the exit? | The group owns Ads, GA4, Search Console, the profiles, and the domain. The firm is a user. The exit, including what is exported and which users come off, is written before the pilot starts. | "We will transfer ownership if you stay," or logins that only the firm can reset. |
| 6. What will you do with a ranking claim or a lead guarantee? | Neither is a deliverable. The firm can show the sheet, the date filter, and the campaign. Google's hiring page says no one can guarantee a number one ranking. | A number one ranking, a guaranteed lead count, or a special relationship with Google. |
Checklist
Twelve checks before you treat the pilot as a decision. Ticks stay in this browser only.
Pilot office checklist
12 items
What DDS Web Solutions builds
DDS Web Solutions builds practice websites. Our team has been building websites since 2014, and we have served 100+ practices. Offices are in Sacramento and San Mateo. The products are Smyly, SmileTrak, and ToothRank. The planning page is DSO and multi-location marketing.
Sources
Wording on these pages can change, so confirm the live page. This is marketing guidance, not legal advice.
- Google Ads Help: Set up a custom experiment
- Google Ads Help: Test with confidence with the Experiments page
- Google Ads Help: About bid strategy statuses
- Google Ads Help: About Smart Bidding
- Google Ads Help: About phone call conversion tracking
- Google Ads Help: About call assets
- Google Ads Help: About location assets
- Google Ads Help: About average daily budgets
- Analytics Help: Apply comparisons to reports
- Analytics Help: URL builders, collect campaign data with custom URLs
- Search Console Help: Performance report (Search results)
- Google Business Profile Help: Understand your Business Profile performance and insights
- Google Search Central: Do you need an SEO?