🔀 Path to Conversion by Device (deprecated)

Dashboard link: https://app.emax-digital.com/path-to-conversion-from-amc

Overview

The Path to Conversion by Device dashboard uses Amazon Marketing Cloud (AMC) data to show which combinations of devices shoppers are exposed to your DSP ads on before they convert, and which device they finally purchased on. Instead of treating each impression in isolation, AMC stitches together the full sequence of devices a shopper encountered — for example "Mobile → Mobile → Desktop → Purchase on Desktop" — and aggregates these into device-path patterns with associated impressions, conversions, sales, and ROAS.

This is particularly useful for understanding cross-device behaviour in DSP campaigns: shoppers often discover products on mobile but complete purchases on desktop or connected TV-adjacent devices, and standard last-click reporting hides this. The dashboard helps you answer: "On which devices should I be spending more impressions to drive conversions on the device shoppers actually buy on?"

This dashboard requires an active Amazon Marketing Cloud (AMC) integration and is only meaningful for brands running Amazon DSP advertising. Sponsored Ads-only accounts will not see results here.

What You Can See on This Dashboard

Section: DSP Path to Conversion by Device

The main table of the dashboard. Each row represents a distinct device path — i.e. an ordered sequence of devices on which a shopper saw your DSP ads before converting — with all the impression, conversion, and revenue metrics for that path.

Key metrics:

• D 1 through D 10+ — The devices in the path, in order. D 1 is the device of first exposure, D 2 the second, and so on up to D 10+ which groups together all longer paths. Devices are categorised by AMC (e.g. Mobile, Desktop, Tablet, Connected TV, Unknown).

• Path Occurrences — Number of unique users who followed this exact device path.

• Impressions — Total DSP ad impressions delivered along this path.

• Conversions — Number of purchase conversions attributed to this path.

• Sales — Total attributed sales revenue (promoted ASINs) from this path.

• Sales Brand — Total attributed brand-halo sales (any ASIN of the advertised brand, not just the promoted ones).

• ROAS — Return on Ad Spend for promoted-product sales (Sales ÷ ad cost).

• Brand ROAS — Return on Ad Spend including brand-halo sales (Sales Brand ÷ ad cost).

Visualizations: A sortable table with one row per device path. Sort by Path Occurrences to find the most common journeys, or by ROAS / Brand ROAS to find the most efficient ones.

Section: Where do users convert?

A simple, high-impact view of the last device in each conversion path — i.e. the device the purchase happened on.

Key metrics:

• Last Device — The final device in the conversion path (Mobile, Desktop, Tablet, Connected TV, Unknown).

• Purchases — Number of purchases that closed on that device.

Visualizations: A pie chart showing the share of purchases by last-touch device, aggregated across all DSP paths for the selected period.

Available Filters

Filter interaction notes

• Reporting Range and Report Date work together — switching from Weekly to Monthly will reload the available Report Date values.

• All three filters apply to both the device-path table and the "Where do users convert?" pie chart simultaneously.

Use Cases

1. Identify the dominant single-device path — Sort the main table by Path Occurrences and look at the top rows. If "Mobile only" or "Desktop only" dominates, your audience is largely single-device and frequency capping per device makes sense. If multi-device paths dominate, cross-device targeting strategies become more important.

2. Find the highest-ROAS device combinations — Sort by ROAS or Brand ROAS to see which device sequences convert most efficiently. A common pattern is that Mobile → Desktop paths outperform Desktop-only paths because mobile drives discovery while desktop closes the purchase — knowing this justifies sustaining mobile impression budget even if mobile alone shows weaker last-click attribution.

3. Validate your last-touch attribution assumption — Open the "Where do users convert?" pie chart. If 70% of purchases close on Desktop but the main table shows most impressions happen on Mobile, last-click reporting will systematically under-credit mobile spend. Use this to defend mobile budget to internal stakeholders.

4. Spot wasted reach — If you see paths with very high impression counts (e.g. D 1 through D 10+ all filled in) but few Conversions and low ROAS, those over-served cross-device users are absorbing budget without converting. Pair this with the Frequency Optimisation dashboard to tighten frequency caps.

5. Brand halo vs. direct sales check — Compare ROAS and Brand ROAS for the top paths. A large gap between the two means a path is driving meaningful brand-level discovery (sales of other ASINs in your portfolio) even when promoted-product ROAS looks modest.

Limitations & Notes

• DSP only. This dashboard is only meaningful for brands running Amazon DSP campaigns. Sponsored Ads impressions are not included in the path construction.

• 14-day data lag: Data is fetched 14 days after each reporting period's end to allow Amazon's standard 14-day attribution window to complete. Weekly data for a given week is available on day 15 after that week closes; monthly data on day 15 after the month closes.

• Device categorisation by AMC. Devices are classified by AMC into a fixed set of categories (Mobile, Desktop, Tablet, Connected TV, Unknown). The "Unknown" bucket can appear for impressions where AMC could not reliably identify the device type.

• Path length cap. Paths longer than 10 device touches are grouped into the D 10+ column; the dashboard does not show the full sequence beyond the first 10 devices.

• AMC privacy thresholds: Device paths followed by very few users may be suppressed or aggregated to comply with AMC's data anonymisation requirements. Very rare path combinations may therefore not appear in the table.

• Attribution scope. Conversions and Sales reflect AMC-attributed purchases within the standard attribution window — these figures may not exactly match the headline numbers in the DSP console or Vendor Central retail reporting.

• AMC integration required. This dashboard depends on an active AMC connection. If data is missing, contact your emax digital account manager.

Data Refresh

Data updates daily, with a 14-day lag after each reporting period ends. For weekly data, complete results for a given week are available on day 15 after that week closes; the same applies for monthly periods.