Searchlight is Aztek's marketing news roundup that brings together the week’s most relevant developments in marketing, search, AI, and digital strategy, all in one place. We update this article throughout the week with news we think is worth your time, along with context to help you understand what changed, why it matters, and what it could mean for your business.
This week's topics:
07/21: ChatGPT Ads: Big Ambitions, Bigger Questions for Marketers
ChatGPT ads are quickly evolving from an experiment into a legitimate media channel, but they’re still far from a must-have in most marketing budgets. The core appeal makes sense. Ads appear directly inside conversations where users may already be researching products or comparing options.
That access to high-intent moments could become valuable over time. For now, though, marketers need to weigh that potential against limited targeting, incomplete measurement, and ongoing brand safety concerns.
What ChatGPT Ads Look Like
ChatGPT ads show up as sponsored answers or small cards beneath the chatbot’s response. The platform limits placements to one ad per conversational turn, creating a cleaner experience than traditional search or display environments.
The format launched earlier this year and is now available in several countries, with a self-service Ads Manager for U.S. advertisers. From a buying perspective, it feels closer to paid search than social media.
The opportunity lies in timing. Users often come to ChatGPT with specific questions or problems, meaning ads can reach them earlier in the research process. While this doesn’t replace search advertising, it introduces another place where high-intent discovery can happen.
Big Revenue Goals, Early-Stage Reality
OpenAI has reportedly set ambitious revenue targets for ChatGPT ads, projecting billions in near-term revenue and much larger growth over time. External forecasts are far more conservative, reflecting how early the market still is. Audience size alone doesn’t make a mature ad platform. Marketers also need reliable targeting, clear reporting, and confidence in where their ads appear. Those systems are still developing.
Early Results Show Promise
Some early advertisers have reported strong click-through rates and efficient acquisition costs, with a few claiming performance that rivals or exceeds display campaigns. These results are encouraging but should be viewed cautiously. ChatGPT’s focused environment may naturally drive higher engagement, and comparisons to display don’t fully capture differences in campaign goals or audience targeting.
Key Limitations to Know
Despite early momentum, several gaps remain:
- Limited targeting: There are no lookalike audiences or detailed demographic filters, and custom audiences require large datasets. This makes the platform harder to use for smaller or niche advertisers.
- Incomplete measurement: New tools like a tracking pixel and Conversions API improve visibility, but reporting is still aggregated. Advertisers can’t see the exact queries that triggered ads or fully map the user journey.
- Brand safety concerns: While keyword exclusions exist, advertisers don’t have full control over conversational context, which can be a concern for regulated or risk-sensitive brands.
What Needs to Improve
For ChatGPT ads to become a core channel, three areas need to mature:
- Targeting: More flexible audience options and lower thresholds for custom audiences.
- Attribution: Better insight into assisted conversions and incremental impact.
- Brand safety: Greater transparency and control over where ads appear.
Until then, most marketers should treat ChatGPT ads as a test channel rather than a primary investment.
What Marketers Should Watch
ChatGPT ads offer a clean user experience and access to potentially valuable intent, but the platform is still in its early stages. Marketers should focus less on bold revenue projections and more on practical improvements in targeting, attribution, and control.
At the same time, brands can prepare by strengthening their own assets. Clear messaging, helpful landing pages, and structured product information will matter more as conversational advertising evolves.
The bigger shift isn’t just about ChatGPT. It’s that conversational platforms are starting to monetize the questions people ask while making decisions. For now, ChatGPT ads are best viewed as a promising experiment that is worth testing, but not yet ready to replace proven channels.
07/23: Instagram Reels Ranking in 2026: Why Watch Time and DM Shares Matter More
Views and likes show initial interest, but they don’t explain why Instagram keeps distributing a Reel. What matters more is what people do after they see it.
Instagram head Adam Mosseri has highlighted three key metrics:
- Average watch time
- Likes per reach
- Sends per reach
Watch time shows if the content holds attention. Sends signal whether it’s worth sharing.
Instagram Is Looking Beyond Visible Engagement
Instagram hasn’t introduced a brand-new algorithm, but its guidance makes one thing clear: not all engagement is equal. Likes still matter, especially with existing followers, but private shares carry more weight because they show intentional value. AKA, someone thought the content was worth sending directly to another person.
That aligns with user behavior. Messaging is now the primary way people share content on Instagram, with billions of Reels reshared daily. The TL;DR: a Reel can get plenty of likes without earning broader distribution.
Why DM Shares Expand Reach
Liking is quick. Sending takes effort. That extra step makes sends-per-reach a strong signal of relevance beyond the current viewer. Mosseri has noted that sends are especially important for reaching non-followers.
Shareable content doesn’t need to be viral or entertaining. It just needs to be useful or relatable—something that solves a problem or captures a shared experience.
Where Saves Fit
Saves aren’t one of Instagram’s top stated ranking signals, but they still matter. People save content they want to revisit, whether that be tutorials, comparisons, checklists, etc. A high save rate suggests lasting value, not just quick engagement. For educational or B2B brands, this can be especially important.
Create Reels Worth Sharing or Saving
You don’t need to force engagement prompts. Instead, build shareability into the content:
- Start with value. Hook viewers immediately.
- Solve a clear problem. Focus on one specific need.
- Make it reusable. Give viewers something worth revisiting.
- Create natural share moments. Content should remind viewers of someone else.
Loops can help with watch time, but clarity and usefulness matter more.
Measure What Matters
Don’t rely on totals alone. Look at engagement relative to reach:
| Traditional Metric |
Better Metric |
| Total views |
Average watch time |
| Total likes |
Likes per reach |
| Total shares |
Sends per reach |
| Follower growth |
Non-follower reach |
| Total saves |
Saves per reach |
There are no universal benchmarks, so compare performance against your own past results.
Test for Real Engagement
Review which Reels drive higher watch time, sends, or saves. Look for patterns in topics and structure. Then iterate. Improve the hook, tighten the message, or refine the example. Don’t rely on reposting trends alone.
Strong organic performance can also inform paid creative, even if it doesn’t guarantee results.
Focus on Value, Not Just Visibility
Watch time and private sharing reinforce a simple idea: visibility isn’t the same as value. Reels perform best when people watch closely, save them, or send them to others. Instead of chasing metrics, create content people actually want to pass along.
07/24: AI Influencers on TikTok Shop: What Brands Gain and What Trust They Risk
TikTok Shop works because people trust people. AI influencers challenge that by removing the human element. Synthetic creators can quickly produce product demos, testimonials, and shoppable videos at scale, and some sellers are already generating hundreds of videos for the same products using AI tools.
The upside? Sure, you can pump out more content faster. The impact on trust? That is less obvious.
AI Makes TikTok Shop Content Easier to Scale
Early TikTok Shop success came from creators testing formats and learning what drove sales. That process took time and resources. AI compresses that timeline. Sellers can generate multiple hooks, swap avatars, tweak voiceovers, and publish variations in minutes.
This is especially useful during peak shopping periods when brands need volume quickly, but it also means users may see many near-identical videos from seemingly different creators.
The Business Case Goes Beyond Cost
AI reduces production friction. No shipping products, coordinating creators, or reshooting content. Brands also gain tighter control over messaging. That control can backfire, though. Real creator content works because it feels lived-in. It's specific, imperfect, and personal. Highly polished, repetitive AI videos can feel generic and easy to scroll past.
TL;DR: Efficiency doesn’t always translate to effectiveness.
Synthetic Testimonials Create a Trust Problem
TikTok is already working to detect and label AI content, and regulators are paying closer attention to synthetic endorsements. The biggest risk is misleading claims. An AI avatar shouldn’t imply personal use or experience. Even accurate product claims can become deceptive when framed as a testimonial.
Disclosure helps, but it’s not enough on its own. Brands still need to evaluate how the message will be interpreted. Before publishing:
- Clearly label AI-generated content
- Avoid claims of personal use or experience
- Verify all product claims
- Document tools and source assets
- Follow platform and regulatory guidelines
- Test Beyond Sales Volume
AI content shouldn’t be deployed at scale without testing. Start small. Compare AI videos with human creator content. Look beyond clicks and GMV: comments, returns, and repeat purchases can reveal trust issues.
Most importantly, don't forget that human creators still matter. As AI content increases, authentic voices may stand out even more.
Brands Still Have to Earn the Recommendation
AI can speed up production and testing, but it can’t replicate genuine credibility. The goal isn’t choosing between AI and human creators; it’s using each where they work best. AI supports scale and experimentation. Humans provide trust and perspective.
