Most guides to marketing measurement assume you have a data team, a warehouse, and a budget for expensive tools. Most marketing teams don't have any of those. If you're a team of two marketers running paid acquisition for a $5M-$20M business, this is for you.
Good measurement doesn't require a data scientist. It requires the right questions, a few reliable data sources, and discipline in tracking what actually matters.
Small teams can build a credible measurement practice with three lightweight tools
Start with what you already have
Your payment processor, Shopify account, or CRM has the ground truth. These systems know exactly how much revenue you generated last week, regardless of which ad platform claims credit for it.
Weekly revenue divided by weekly ad spend gives you your blended marketing efficiency ratio (MER). This calculation is free, takes 10 minutes per week to maintain, and is more reliable than any attribution dashboard. It doesn't tell you which channel is working, but it tells you whether your overall advertising investment is generating business value. Track it in a spreadsheet every week.
If MER is stable or improving as you scale spend, the overall program is working. If MER is declining, something is wrong, either the advertising is becoming less effective or costs are rising faster than revenue. This is the single most useful number most small marketing teams aren't tracking.
Fix your GA4 before buying anything else
Most GA4 implementations are misconfigured. The most common problems:
Purchase events aren't firing correctly or are double-counting. Check this by comparing GA4 purchase revenue to your Shopify or CRM revenue for the same period. If they're more than 10-15% apart (in either direction), something is broken.
Internal traffic isn't filtered. If your team visits your site regularly and those sessions are counted as users, your engagement metrics are inflated and your conversion rate is understated. Create an IP filter or internal traffic definition in GA4 to exclude your office and any remote workers' regular IPs.
Cross-domain tracking is broken if you have separate domains for checkout and your main site. Customers who add to cart on shop.yourdomain.com and check out on checkout.yourdomain.com will appear as two separate sessions, and the conversion often gets attributed to "direct" instead of the actual acquisition channel.
Fix these three issues and you have materially better attribution data at zero additional cost. GA4's cross-channel view isn't perfect, but a properly configured GA4 is more useful than an expensive MTA tool on top of broken tracking.
The simplest incrementality test you can run
Take your largest retargeting campaign. Identify the audience it's targeting (site visitors, cart abandoners, or product viewers). In your ad platform, create an audience of people who qualify for retargeting but are explicitly excluded from seeing your retargeting ads. Track both groups in your backend for four weeks.
After four weeks, compare the conversion rate of the retargeted group and the non-retargeted group. Adjust for any differences in recency or intent level between the two groups. The gap is a rough estimate of how much your retargeting is actually causing conversions rather than just claiming credit for customers who would have bought anyway.
This test doesn't require a holdout test platform or data engineering infrastructure. The setup takes a few hours and the analysis takes a spreadsheet. The result will often surprise you: retargeting incrementality is consistently lower than attributed ROAS implies.
A manual measurement tracker
Build one spreadsheet that you update weekly. Track:
- Total revenue (from your backend)
- Ad spend per channel (from each platform dashboard)
- MER (total revenue / total spend)
- Spend share per channel (what percentage of your total spend is going to each platform)
- Week-over-week MER change
This is not a sophisticated analytics system. It's a scorecard. But it forces you to look at backend revenue every week instead of only looking at platform dashboards. Over time, you'll start to see patterns: what spend levels and channel mixes correlate with strong MER, when MER degrades, and whether channel mix shifts improve or hurt efficiency.
Add one more column when you run any test: a note on what changed that week. If you turned off retargeting to run an incrementality test, note it. If you launched a new prospecting campaign, note it. Annotating your data as you collect it is the lowest-effort way to make the data interpretable months later.
Open-source incrementality tools
Meta's GeoLift is a free package for running geo-based incrementality experiments. It runs in R or Python and walks you through market selection, test design, and result analysis. You don't need a data scientist to run it if you're comfortable following documentation and running basic scripts. Budget 4-6 hours to set up your first test.
Google's Conversion Lift and Brand Lift tools are built directly into Google Ads at no additional cost. For brands spending meaningfully on YouTube or Display, these give you an incrementality measurement without any external tools.
Facebook's Conversion Lift is available in Meta Ads Manager as a native feature. It runs a holdout test within Meta's platform automatically. The limitation is that it only measures Meta's incrementality within Meta, not against your backend data, so the results should be treated as directional rather than precise.
For the first year of measurement without a data team, these free native tools plus your weekly MER tracker cover most of what you need.
What to ignore at this stage
MMM requires at minimum 12-18 months of consistent multi-channel spend data, variation in that spend week over week, and enough statistical signal to distinguish channel effects from seasonality. Sub-$100k/month budgets rarely produce enough signal for MMM to work reliably. The open-source tools are free, but the time required to run and interpret them correctly isn't worth the investment at this scale.
Expensive MTA tools like Northbeam or Rockerbox are priced for brands spending $200k+/month on paid. Below that threshold, the decision quality improvement from these tools doesn't justify the subscription cost. GA4 properly configured does 70% of what they do at zero cost.
Complex attribution modeling, data warehouses, and BI tool subscriptions are the right investment when you have a data team to maintain them and business decisions that require their output. Without a data team, the tooling becomes a cost without a return.
When to bring in help
The point where DIY measurement breaks down is when you're spending more than $200k/month and facing budget allocation decisions worth hundreds of thousands of dollars annually. At that scale, the marginal value of better measurement is high enough to justify external help.
A good freelance analytics consultant (or a measurement-focused agency) can set up incrementality testing infrastructure, run your first MMM, and build a measurement framework that your marketing team can operate independently. That engagement typically costs $10,000-$30,000. At $200k+/month in ad spend, a 5% improvement in allocation efficiency more than pays for it.
Before that scale, the investment in tools and consultants often exceeds what better measurement could actually save you.