Soku AIBlog
Insights, tutorials, and updates from the Soku AI team.
Best DSP Platforms for Advertisers in 2026: What Each One Actually Gives You Access To
Almost no demand-side platform publishes what it costs, so comparing DSPs on price is a dead end. Compare them on what they let you buy instead. A platform-by-platform guide to Amazon DSP, DV360, The Trade Desk, StackAdapt, Yahoo, Microsoft Invest, Viant, Basis and Adform — self-serve versus managed, the one published spend minimum in the category, and the spend level where each starts to make sense. Verified 2026-09-22.
14 min read · September 22, 2026
Best CTV Advertising Platforms, Ranked by How Easily You Can Actually Start (2026)
A ranking of connected TV advertising platforms in 2026 built around the question the listicles avoid: what is the real entry point? Verified published minimums, self-serve versus managed routes, measurement realities and the creative work CTV quietly demands.
13 min read · September 21, 2026
Best Programmatic Advertising Tools: The Full Stack, Layer by Layer (2026)
Programmatic advertising is not one tool, it is seven layers. A practical map of the DSP, SSP, ad server, identity, verification, creative and measurement layers in 2026 — which ones you actually need at your spend level, what each costs, and where teams over-buy.
14 min read · September 21, 2026
Retail Media Networks, Explained for Advertisers (2026)
What a retail media network actually is, why two companies take almost all of the growth, the difference between on-site, off-site and in-store inventory, how the measurement is genuinely better and quietly self-marked, and how to decide whether an RMN belongs in your plan.
13 min read · September 21, 2026
ChatGPT Ads Policy v1.6: OpenAI Can Now Decline Advertisers Over Its Own Competitive Position
On 2026-09-10 OpenAI published v1.6 of its Ad Policies. The discretionary-refusal clause grew from refusing ad content to declining, restricting or removing advertisers — and now names business interests and competitive position as grounds. The exact before-and-after wording, who is actually exposed, what the documentation still does not say, and the four developer-doc changes that shipped in the same window.
13 min read · September 14, 2026
ChatGPT Ads Platform Targeting Comes to the API: targeting.platforms, the Six Values, and the Bulk Gap That Stayed Open
OpenAI split the ChatGPT Ads Campaign Targeting doc into three and shipped targeting.platforms.included to the Advertiser API, with six accepted values and explicit clearing semantics — null clears, empty objects return HTTP 400. The Bulk API still has no platform field. Exact before-and-after wording from both documentation surfaces, verified 2026-09-11.
12 min read · September 11, 2026
ChatGPT Ads Now Paces Campaign Total Budgets: What Changed, and the 365-Day and 60-Day Pacing Periods
OpenAI rewrote the ChatGPT Ads budget documentation and published a new Budget Pacing article. The campaign total budget went from an undescribed spending ceiling to a paced budget type, with a 365-day maximum pacing period, a 60-day default when there is no end date, and an explicit rule for what happens when you change the budget mid-flight. The exact before-and-after wording, verified 2026-09-09.
11 min read · September 9, 2026
ChatGPT Ads Event Quality: What the Score Measures and How to Fix Every Warning
OpenAI shipped an Event Quality score for ChatGPT Ads data sources — nine named warnings, a seven-day rolling window and four separate reasons a score can simply be unavailable. This is what each warning actually checks, which field fixes it in the Pixel or the Conversions API, and why the score is a setup diagnostic rather than a performance forecast. Verified against the official documentation on 2026-09-07.
12 min read · September 7, 2026
ChatGPT Ads Identifier Normalization: OpenAI Rewrote the Hashing Rules, and Leading Zeroes Now Matter
OpenAI rewrote the normalization rules for hashed identifiers in the ChatGPT Ads Measurement Pixel and Conversions API. The phone rule changed from removing a leading + or 00 to removing a leading + and any leading zeroes, which changes the hash for every number stored in national format. The exact previous and current wording, a worked example table, the accent rule almost every implementation gets wrong, and what the help centre still does not say. Verified 2026-08-31.
12 min read · August 31, 2026
Five Checks Before You Let an Agent Spend Your Ad Budget
Autonomy is not a slider you turn up. Five specific checks — attribution of cause, hard floors, decision replay, missing-data behaviour, and rollback ownership — that decide whether an agent managing your campaigns compounds your results or your mistakes.
8 min read · August 31, 2026
Your Ad Agent Can't See the Auction It's Bidding Into
A rising CPC caused by creative fatigue looks identical to a rising CPC caused by a new bidder. We measured two real competitive sets to show why an ad agent optimizing inside one account cannot tell them apart.
9 min read · August 28, 2026
AI vs Human Campaign Audit Accuracy: A Reproducible Test
Everyone claims AI audits ad accounts as well as a specialist. Almost nobody publishes a protocol you could use to check. Here is the full test design — the seeded-defect method, the scoring rubric, the traps that make naive comparisons meaningless — so you can run it on your own account instead of trusting a vendor's number. Verified 2026-08-20.
16 min read · August 20, 2026