Last-touch attribution assigns 100% of the credit for a conversion to the last measurable interaction before that conversion occurred. In a click-based digital world, that last interaction is almost always a search ad click or a direct visit to the site. CTV, which has no click, gets nothing. The practical effect: brands using last-touch attribution to evaluate their CTV campaigns will consistently conclude that TV contributed little, cut the budget, and redirect it to search — which was simply the last step in a journey that television initiated.
This article explains how last-touch works, why it fails specifically for CTV, what it misattributes credit to, and what alternatives are worth using instead.
How last-touch attribution works
Last-touch is the simplest attribution model. When a conversion event fires — a purchase, a sign-up, an install — the measurement system looks backward through the conversion path and finds the most recent tracked touchpoint. That touchpoint gets full credit. Everything before it gets zero.
In a typical conversion path for a TV-influenced buyer: the person sees a CTV ad (not tracked with a click), does nothing immediately, searches the brand name three days later, clicks the search ad, and converts. Last-touch gives the search ad 100% credit. The TV ad gets 0%.
This model made sense when advertising was primarily search and display — channels where every interaction leaves a click trail. It has never made sense for channels with no direct click mechanism: radio, outdoor, linear TV, and CTV.
Why last-touch systematically undervalues CTV
CTV fails last-touch attribution for three structural reasons:
No click mechanism
CTV ads are non-interactive in the standard viewing environment. There is no click-through. The viewer watches the ad on the television screen and then — if the ad worked — takes action on another device, often minutes or hours or days later. Without a click event, last-touch models have no record of the TV exposure. The impression was served, the ad was watched, but the attribution system never saw it happen.
Long consideration cycles
CTV is primarily an upper- to mid-funnel medium. It builds awareness and consideration. The conversion happens at the end of a consideration process that might last days or weeks. Last-touch attribution windows — typically 7 or 28 days in most ad platforms — may not even capture the full journey, let alone the CTV touchpoint at the beginning of it. A vehicle purchase influenced by automotive CTV advertising might convert 60 days after the initial exposure. No standard attribution window covers this.
The shared screen problem
The TV is a household device. Attribution systems cannot reliably connect the individual who watched the TV ad to the individual who later clicked on a search result from their mobile phone. Even if the exposure was logged, matching the TV viewer to the converter requires a device graph that works at household level — which most attribution platforms do not fully support in India.
What last-touch misattributes credit to
When CTV gets zero, something else gets 100%. The beneficiaries of last-touch misattribution are predictable:
- Branded search: A viewer who searches a brand name after seeing the CTV ad is logged as a search conversion. Search gets the credit. The CTV ad that prompted the search is invisible. Branded search volume is often a proxy signal for upper-funnel TV effectiveness — but last-touch inverts this by rewarding search at TV's expense.
- Direct traffic: Viewers who type the URL directly after seeing the ad generate direct conversions. Again, these are logged as direct, not TV.
- Retargeting: If a viewer who saw the CTV ad was also in a retargeting audience and saw a display ad before converting, the retargeting display ad takes the credit. The CTV ad that put the viewer in the consideration set is not credited.
The real cost of last-touch for CTV budget decisions
The measurement error has a direct budget consequence. Brands that evaluate channel performance using last-touch will see CTV showing near-zero direct attribution, search showing high attribution, and retargeting appearing highly efficient. The natural response is to shift budget from CTV to search and retargeting. But this decision is based on faulty measurement, not faulty performance. The search and retargeting channels are harvesting demand that CTV created — they look efficient precisely because the demand was already there. Cut the CTV spend and the search and retargeting performance degrades, because there is less demand being created at the top of the funnel.
This is the flywheel problem: last-touch causes under-investment in CTV, which reduces upper-funnel demand creation, which reduces branded search volume, which eventually reduces conversion volume — but by then, attribution has moved on to other explanations.
When last-touch is acceptable
Last-touch is not useless in every context. It can work when:
- The entire purchase journey happens in a single session — high-intent search clicks for commoditised products where consideration is minimal.
- You are measuring search and direct channels in isolation, not comparing across channels.
- You have no alternative and are using last-touch as a directional baseline while building better measurement infrastructure.
For CTV specifically: last-touch is never appropriate as the primary measurement methodology. If you must use it for reporting, treat CTV numbers as a floor, not a ceiling.
What to use instead of last-touch for CTV
Three alternatives are more appropriate for CTV measurement:
Incrementality testing
Create a holdout group excluded from CTV targeting. Compare conversion rates between the exposed and holdout groups. The difference is the causal lift from your CTV campaign — not a model-assigned share of credit, but an actual measurement of what the campaign caused. This is the most reliable approach for CTV and does not require cross-device identity resolution.
Branded search uplift
A simpler directional measure: track branded search volume before, during, and after a CTV campaign. CTV ads consistently drive branded search uplift that is measurable in Google Search Console or a keyword tracking tool. It is not full attribution, but it is a real signal that CTV is driving consideration. India-specific note: branded search uplift measurement is particularly accessible because Google dominates search in India and the tools are widely available.
Media mix modelling
MMM uses aggregate spend and sales data to estimate channel contribution. It does not require individual-level tracking. For brands with large enough budgets and consistent long-term data, MMM gives a more defensible estimate of CTV's contribution to the overall media mix than any click-based model can provide.
India context: last-touch in the Indian advertiser landscape
Last-touch attribution is widely used by Indian advertisers because it is the default in most ad platforms — Google Ads, Meta, and most DSPs default to last-click models. Indian CTV budgets are frequently evaluated using these platform-native attribution tools, which consistently underreport CTV performance. The result is a structural underinvestment in CTV relative to its actual contribution to the purchase journey.
Indian advertisers who have tested CTV with incrementality testing consistently find that CTV's true contribution is higher than last-touch numbers suggest — often 2x to 4x higher for categories like e-commerce, BFSI, and consumer durables. The measurement gap is not a CTV performance problem; it is an attribution methodology problem.