ML Ads Manager

01 Overview

Mercado Libre's advertising product Mercado Ads had generated strong adoption among enterprise sellers with dedicated marketing teams. But 72% of sellers on the platform were SMBs running their business as solo operators or with teams of 2 3 people. This segment had almost zero adoption of the ads product, not because they didn't want to advertise, but because the interface assumed advertising expertise they didn't have.

The ML Ads Manager project was a complete redesign of the self-serve ads interface for the SMB seller segment building a guided, intelligent campaign creation flow that removed every expert assumption from the process while preserving full control for sellers who wanted it.

02 The Problem

The existing Mercado Ads interface opened to a dashboard of campaign performance metrics. For sellers who had never run a campaign, this was a blank screen with no obvious action. Creating a new campaign required selecting a campaign type (Sponsored Products, Sponsored Brands, or Display), a bidding strategy (CPC, CPM, or target ROAS), a daily budget, a set of keywords, and a landing destination all before seeing any estimate of what the campaign would actually do.

Research with 30 SMB sellers who had never used Mercado Ads showed a consistent failure pattern: they would get through campaign type selection, stall at bidding strategy (not knowing what CPC or ROAS meant), and abandon the flow entirely. The exit point was always the same screen the one that required the most advertising expertise to complete.

03 Process

The redesign principle was "goal-first, settings-later." Instead of asking sellers to configure the mechanism of the campaign (bidding strategy, keyword match type, placement targeting), the interface first asks them what they want to achieve: more views, more sales, or more visibility for a new product. The platform then configures the campaign settings automatically based on the goal selected.

Budget recommendation was rebuilt using an AI model that analyzes the seller's product category, competitive density, average selling price, and typical margin to suggest a starting budget with a projected outcome range. Sellers see "invest $50/week, estimated 200 400 additional product views" rather than a blank budget input field.

04 Solution

The ML Ads Manager opens with a smart campaign wizard that guides the seller through 4 questions: What do you want to promote? What's your goal? What's your budget? When do you want to start? The platform handles everything else keyword selection, bid calibration, placement configuration automatically. The whole flow takes under 5 minutes.

Campaign performance is presented as a plain-language summary alongside the metrics: "Your ad reached 1,240 potential buyers this week. Sales from this campaign: 8 units, $4,200 revenue. Your ad spend this week: $48." For sellers who want to go deeper, a full analytics view is one click away, but the plain-language summary is the default because it's what 85% of SMB sellers actually need to make decisions.

05 Results