Updated September 24, 2026 · By Emiliano Ubaldi
The least expensive campaign mistake is often the one you catch before launch. AI can help examine an idea early, but the most valuable pre-launch work still includes straightforward checks: the offer is accurate, the page works and the conversion event measures the right action.
Treat waste estimates as a prompt to investigate
Respondents in Smartly’s 2026 research estimated that about 20% of their annual digital marketing spend was wasted on average. That is a self-reported estimate, not an audited loss rate that applies to every advertiser. The useful response is to identify where waste might occur in your own journey.
For a dealership, possible problems include ads for unavailable vehicles, landing pages that omit important conditions, duplicated lead events or slow follow-up. None requires a sophisticated predictive model to become expensive.
Use synthetic feedback to form better questions
Synthetic audiences use models to simulate reactions rather than recruiting actual customers. They can help explore whether a message is confusing, compare alternative explanations or surface assumptions worth testing. Predictive scoring may also help prioritise which creative ideas deserve a closer look.
These outputs are hypotheses. A plausible simulated response is not evidence of actual buying behaviour, and repeated agreement from the same model is not independent validation. Results depend on the underlying data, instructions and representation of the audience.
Build a practical pre-flight check
Before approving spend, verify the destination on mobile, submit an authorised test enquiry where appropriate, confirm the event fires once and check that it reaches the intended team. Review geographic targeting, exclusions, current inventory, contact details and any offer restrictions.
Then give a small set of clearly different ideas a controlled live test. Decide the question, budget and success measure in advance. Avoid changing multiple important variables at once if the goal is to understand why performance differs.
For lead generation, I would look beyond click-through rate to valid enquiries, qualification and appointments. Allow for the time it takes the sales team to work a lead. A campaign that generates cheap submissions can look efficient before the quality information arrives.
Use what you learn to refine the next test. AI earns its place when it helps the team make a better decision—not when it produces a confident score that nobody can challenge.
Explore a more deliberate approach to paid-media lead generation.
Illustration generated with AI for IM Advise.