Marketing Analytics Stack for 2026: Small Business and Solo Media Buyer

By Jan 29, 2019 10 min read Updated: Sep 8, 2026

Build Your First AI Project This Weekend

Stop consuming tutorials. Start creating. Get the free step-by-step guide.

Stop consuming tutorials. Start creating. Get the free step-by-step guide.

2026 Update:

This used to be a 2019 affiliate-tracker roundup. Voluum vs Binom vs whoever was cheapest that month.

That is not the stack most of you need.

If you run a local business (HVAC, dentist, lawyer, plumber), your problem is: which Google Ads keyword, GBP action, and landing page produced the call. You need Plausible or GA4 on the site, CallRail on the phone, and Clarity so you can watch the form that nobody submits.

If you buy media for a living, you still need a click tracker. That is a different article: the click tracking guide.

Two jobs. Two stacks. Do not buy Voluum because a 2016 blog post said “data is the foundation of profits” and then never record a phone call.


You’re spending money. Page is live. Ads are on.

You still cannot answer: what actually made a customer?

Not “sessions.” Not “engaged users.” A customer. A booked job. A qualified call over 90 seconds. A form that turned into a quote.

I’ve watched people install six tools and still guess. I’ve also turned a losing account around with three: a clean site analytics install, call tracking, and session recordings on the landing page.

This guide is the stack I would install this week. Not a catalog of every tracker that ever ran a webinar.


Quick Navigation

SectionWhat You’ll Build
The only numbers that matterWhat to track before you buy software
Small business stackPlausible, CallRail, Clarity, pixels
Solo media buyer stackClick tracker + CAPI + the same site tools
What not to buy yetAttribution suites, ad servers, extra CRMs
AI on top of the dataExport, prompt, act
Setup checklistDo this before you raise budget

The Only Numbers That Matter

Tools are how you collect. They are not the strategy.

Write these down for your business before you open a pricing page:

QuestionLocal serviceMedia buyer / affiliate
Did they show up?Site sessions by source (GBP, Google Ads, organic, referral)Clicks by campaign, placement, creative
Did they do the thing?Calls, forms, chats, booked jobsLeads, sales, deposit, call duration
What did it cost?Cost per booked job, not cost per clickCPA / ROI against real payout
Where did they leak?Bounce on mobile, form field drop, missed callLP CTR, offer CVR, geo/device losers

If you cannot measure the “thing,” more dashboards will not help.

For a dentist, the thing is usually a phone call or an appointment request. For a native buyer, it is a conversion postback. Pretending those are the same stack is how you waste $200/month on a tracker you never log into.


Small Business Stack (Local Service)

This is the default if you generate leads in a city. HVAC. Plumbing. Law. Chiro. Clinic.

1. Site analytics: Plausible (or GA4 if you must)

Plausible is the one I actually look at.

Hosted Starter is $9/month for 10k monthly pageviews, one site, 3 years retention. Growth is $14 (3 sites). Business is $19 (10 sites, funnels, Stats API). 30-day trial, no card. Script is tiny. No cookie banner. UTM campaigns show up without a GTM science project.

Use GA4 anyway if you run Google Ads and you want native audience building and conversion import the way Google prefers it. GA4 is free. It is also noisy. I treat it as the Google-ads companion, not the dashboard I live in.

Do not make GA4 your only view of the site. Consent mode, ad blockers, and 14-month default event retention on the free property mean you will argue with incomplete numbers.

Take Away: Plausible for “what happened on my pages.” GA4 for “Google Ads needs a conversion.” Both can coexist.

2. Phone: CallRail

If the business lives on inbound calls and you skip call tracking, you will lose the argument with the owner. They will say the leads were junk. You will have no recording.

CallRail is the default. Base Call Tracking starts around $45/month with 5 local numbers and 250 local minutes. Conversation Intelligence is the next rung if you want transcripts and summaries. Extra numbers and minutes are usage (typically about $3/number and $0.05/local minute; confirm on their pricing page).

They sunset Lead Center, the VoIP/softphone, on January 27, 2026. Tracking numbers still forward. You need a real destination phone.

Wire dynamic number insertion on the site. Send call conversions into Google Ads. Then you can see which keyword rang the phone instead of which keyword got a click.

Pair it with missed-call capture so after-hours traffic is not voicemail.

3. Behavior: Microsoft Clarity (free)

Clarity is free heatmaps and session recordings. Watch 10 sessions from paid traffic before you “redesign the hero.” You will find the unclickable button, the form that jumps on iPhone, or the click-to-call number that is not the tracking number.

Hotjar is fine if you need surveys. I start with Clarity because the price is zero and the recordings are enough.

4. Platform pixels

Only the platforms you actually buy:

  • Google Ads tag + Enhanced Conversions
  • Meta Pixel + Conversions API if you run Meta
  • Microsoft UET if you run Microsoft Ads

Server-side (CAPI / Enhanced Conversions) is the 2026 default, not a power-user extra. Browser pixels miss people. How many? Enough that I will not quote a fake “20-40%” without a study I trust for your site. Turn CAPI on. Compare platform-reported conversions to CallRail and to the CRM. That gap is your real number.

5. Email you own: Kit

Traffic you do not capture is rented. Kit (formerly ConvertKit) has a free plan up to 10,000 subscribers. Paid Creator starts at $33/month billed yearly at 1,000 subscribers (kit.com/plans). Use it for the follow-up after the missed call and the quote.

What this costs

NeedToolBallpark
Site analyticsPlausible Starter$9/mo (pricing)
Google Ads companionGA4$0
CallsCallRail base~$45/mo + usage (pricing)
Recordings / heatmapsClarity$0
PixelsNative tags + CAPI$0 software, some setup time
EmailKit free or Creator$0-$33+/mo (plans)

You can be fully instrumented for a local account for about the cost of one wasted Google Ads day.


Solo Media Buyer Stack

You buy native, search, social, or pops. You need to know which placement paid you.

Site analytics still matter. They do not replace a tracker.

1. Click tracker (source of truth for the buy)

A tracker sits between the ad and the lander, stamps a click ID, and takes a postback when the offer converts. Pixels in the browser are the backup, not the system.

Full mechanics: click tracking guide.

Starting points in 2026 (confirm on vendor pages; they move tiers):

TrackerHostingWhy it exists
RedTrackCloudTrial, CAPI, less server babysitting
VoluumCloudTeams, automations, MCP; priced by events
BeMobCloudCheap / free tier to learn
KeitaroSelf-hostFlat license, you own the box
BinomSelf-hostSpeed, unlimited events, you own the box

Rule I use: If monthly ad spend is still in test-budget land, cloud. If you are pushing serious volume and you care who holds the log files, self-host.

Do not install a tracker “because affiliates do” if you only run Google Ads for a dentist. CallRail + Google Ads conversion import covers that.

2. Same site layer as the small business

Plausible or GA4 on the lander. Clarity on the lander. You want to know if the page is broken, not just that CVR is 0.4%.

3. Server-side conversions to the ad platform

Meta CAPI, Google Enhanced Conversions, TikTok Events API. Your tracker should fire these or you should fire them from the server. Last-click in GA4 is not your source of truth when you buy from five networks.

4. Calls, if the offer is a call

Same CallRail (or Ringba if you are routing to buyers). A click tracker that never sees the phone is half-blind in home services.


What Not To Buy Yet

Hyros / Triple Whale / “unified attribution”

Useful when you have multiple paid channels, real revenue, and a person who will log in. Not useful at $2k/month on one channel. Get one channel profitable first.

Your own ad server

Almost nobody reading this needs Revive. Pixels are not an ad server. I wrote the honest 2026 version.

A second CRM “for marketing”

If the office lives in Jobber, Housecall, or HubSpot, send CallRail and form webhooks there. Do not add a fourth inbox.

Enterprise heatmaps, extra session tools, three analytics products

Clarity + Plausible. If you cannot act on that, Mixpanel will not save you.

OpenVBX

Dead. Twilio abandoned it in 2017. Don’t install it.


Cloud Tracker vs Self-Host vs “I Don’t Need One”

SituationDecision
Local service, Google/Meta, phone leadsNo click tracker. CallRail + Plausible + pixels
Content site, SEO, emailPlausible. Kit. No tracker
Affiliate / native / multiple networksCloud tracker first
High volume, you want the data on your VPSKeitaro or Binom + a server you will actually maintain

Self-hosted is not “more pro.” It is more ops. If you will not patch the box, stay on cloud.


AI On Top Of The Data

AI does not replace the stack. It reads the CSV faster than you do.

Weekly, not daily:

  1. Export CallRail (calls, duration, source) or tracker (spend, clicks, CVR, ROI)
  2. Export Plausible (pages, sources, goals)
  3. Paste into Claude with a tight prompt
This is last week's data for [business].

1. Which sources produced the outcomes I care about (calls over 90 seconds / purchases), not just clicks?
2. Which pages or campaigns have traffic and no outcomes?
3. Three cuts I can make this week. Three tests worth more budget.
4. Flag anything that looks like a tracking break (traffic up, outcomes zero).

Do not give me a lecture. Give me a table and next actions.

If you buy media, use the longer prompts in the tracking guide. If you are on Plausible, the weekly workflow is in the Plausible guide.

Voluum now ships an MCP. Kit ships Kit MCP on Creator. Cool if you already live in those tools. Do not pick a tracker because it has an MCP badge.


Setup Checklist

Do this before you raise budget.

Every business:

  • One analytics tool you will actually open (Plausible, plus GA4 if you buy Google)
  • Conversion goal: thank-you URL, form event, or call
  • UTM on every paid URL
  • Clarity recording on the money pages
  • Platform pixel + server-side conversion for each ad account you spend in

If the phone matters:

  • CallRail number on ads and GBP (or DNI on the site)
  • Recording + recording disclosure
  • Call conversions in Google Ads
  • After-hours path (missed calls)

If you buy from affiliate networks or native:

  • Tracker link with click ID
  • Postback tested (one test conversion, both systems agree)
  • Placement / site / creative parameters populated

Missing the phone piece on a call business is the expensive one. Missing the postback on an affiliate campaign is the other expensive one.


Mistakes I Still See

Installing GA4 and calling it done. GA4 does not record the call. It does not tell you which native zone is draining you. It does not replace CallRail or a tracker.

Buying a tracker for a Google Ads / GBP account. Use CallRail and Google’s conversion import.

Skipping recordings. You will not “just remember” why the lander dies on Safari.

Collecting everything, reviewing nothing. Five numbers: spend, outcomes, cost per outcome, conversion rate, and one quality proxy (call length, show rate, refunds). The rest is until you are profitable.

Quoting industry “pixel loss” percentages in client decks. Measure your gap: platform vs CallRail vs CRM.


What To Do This Week

Local business / local ads:

  1. Start Plausible (trial) and put the script in the head. Full steps: Plausible guide.
  2. Start CallRail. One number. Recording on. Setup walkthrough.
  3. Install Clarity on the landing page. Watch 10 paid sessions.
  4. Turn on Enhanced Conversions / CAPI for the ads you actually run.

Solo media buyer:

  1. Read the click tracking guide and run one test click + test postback before spend.
  2. Put Plausible + Clarity on the lander anyway.
  3. If the offer is inbound calls, add CallRail or Ringba. Do not “track the click” and ignore the phone.

The edge is not a longer tool list. It is knowing which $50 came back.

Related:

Questions? Contact me.