The sticker price on your last batch of AI-assisted blog posts was probably somewhere around $200 per draft. Maybe less. That number feels efficient. It's also the wrong number to optimize for, and the March 2026 core update made the gap between "cost per draft" and "cost per asset that actually ranks" wide enough to wreck a quarterly budget.
We've been modeling this internally for months, and the math keeps landing in the same uncomfortable place: for two-person B2B teams spending under $2,000/month on content, the true cost of a post that reaches organic traffic is 3-5x what shows up on the invoice. The gap isn't production. It's indexing survival.
The Draft Price Illusion
AI moved content production costs down significantly. Median costs for AI-augmented teams now sit around $847 per published asset, compared to $2,310 pre-AI. That's a real reduction. Nobody's arguing otherwise.
But "published" and "ranked" are not the same thing. They used to be close enough to treat as interchangeable. A post went live, Google crawled it within days, and it entered the index. From there, it either gained traction or it didn't, but at least it got a shot.
That assumption broke in Q1 2026.
What the March 2026 Core Update Actually Changed for Small Teams
The March 2026 core update rolled out between March 27 and April 8, and it hit harder than most teams expected. More than 55% of monitored websites saw measurable ranking changes, with roughly 24.1% of pages that had been in the top 10 dropping out of the top 100 entirely.
The winners were predictable: official institutions, specialist publishers, and pages with clear original value. Sites closest to the underlying source of information gained. Everyone else lost ground.
Here's the part that matters for budget planning, though. The update didn't just reshuffle rankings. It tightened indexing criteria. To secure a spot in the index now, content needs what Google's systems effectively treat as "net-new value," information that doesn't already exist in the searchable library. Generic posts that rephrase existing knowledge, even well-written ones, are increasingly ignored by the crawler.
This is not a quality filter in the way content marketers typically use that word. A well-structured, grammatically perfect, SEO-optimized post can still fail to index if it adds nothing new. And that's exactly the kind of content most AI-assisted workflows produce at scale.
Why Indexing Survival Is the Variable Nobody Budgets For
Botify's analysis of 6.3 billion URLs across 1,000 enterprise websites found that only 3.5% of pages receive any organic traffic at all. Enterprise sites have different dynamics than small B2B blogs, sure. But the directional signal is consistent: a meaningful share of published content generates zero organic return.
For a two-person team publishing 8-10 posts per month on a $2,000 budget, let's model what happens when indexing survival drops below 100%.
Assume a 65% indexing rate, which is generous given the March 2026 tightening. Out of 10 posts, roughly 6-7 enter the index. Of those, maybe 4-5 generate any measurable organic traffic within 90 days. The rest sit in Google's crawl queue or get indexed but never rank for anything with search volume.
Your $200-per-draft cost just became $400-$500 per asset that generates any organic signal. And we haven't even touched the question of compounding yet.
The Compounding Window Is Now Conditional
Content marketing's economic argument has always rested on compounding: a post published today generates traffic for months or years, unlike paid ads that stop the moment you stop paying. That argument still holds, but only for content that clears two gates.
Gate one: indexing. Gate two: citation.
Google AI Overviews now appear on 48% of all search queries, and organic click-through rates drop by 58-61% when they do. So even indexed, ranking content faces a CTR tax it didn't face 18 months ago.
But there's a split forming. Brands cited inside AI Overviews earn 120% more clicks per impression than uncited competitors, with a 35% traffic lift. The content that gets cited, and the content that doesn't, are diverging into two entirely different economic tiers.
The overlap between "ranking on page one" and "being cited in AI Overviews" has collapsed. Through mid-2025, about 75% of URLs cited in AI Overviews also ranked in the organic top 10. By early 2026, that overlap dropped to 17-38% depending on query class. Ranking and being cited are now separate games with separate rules.
Building the Indexed-Asset Yield Model
We built this model for teams like ours: small, budget-constrained, trying to generate organic pipeline without an agency retainer. The numbers below use conservative assumptions. Your actuals will vary, but the structure holds.
Inputs:
- Monthly content budget: $2,000
- Posts produced per month: 10
- Cost per draft (all-in, including tools and time): $200
Indexing layer:
- Posts indexed within 30 days: 7 (70%)
- Posts generating any organic traffic within 90 days: 5 (50%)
- Cost per traffic-generating asset: $400
Compounding layer:
- Posts still generating traffic at month 6: 3 (30% of total production)
- Posts cited in AI Overviews or similar: 1-2 (10-20% of total production)
- Cost per compounding asset: $667-$1,000
The yield: Your $200 draft is actually a $667-$1,000 ranked asset, assuming you're producing content with enough original value to survive indexing. If your indexing rate is lower (and for teams running pure AI generation without editorial judgment, it often is), multiply accordingly.
This is why cost-per-draft benchmarks are misleading. They measure the input. The output that matters, a post that indexes, ranks, compounds, and potentially gets cited in AI answers, costs 3-5x more per unit than the production sticker price suggests.
The Structural Traffic Problem Underneath
Even the content that clears both gates faces a headwind that didn't exist two years ago. U.S. organic search traffic declined 2.5% year over year in April 2026, marking the first sustained multi-month decline in the history of the modern web. This isn't seasonal. It's structural.
Around 60% of all searches now end without a click to any website. When an AI summary appears, users click a traditional result just 8% of the time, compared to 15% without one.
So the compounding curve itself is flattening. A post that would have generated 500 organic visits per month in 2024 might generate 300 in 2026, all else equal. That doesn't kill the ROI argument for content marketing. It does mean your per-asset economics need to be sharper than ever.
What Actually Survives Indexing (and What Doesn't)
We've watched enough Search Console data across client blogs to see patterns. Posts that survive the post-March 2026 indexing filter tend to share specific characteristics.
Original data performs. Marketers who use original data see 39% more organic traffic. This includes survey results, internal benchmarks, experiment write-ups, and customer-derived insights. If your post is synthesizing information that already exists elsewhere, the indexing probability drops.
Operational specificity matters. Posts that describe real processes, actual failures, specific numbers from real campaigns, those index and rank. Pages with real operational details have an average dwell time 2.3x that of pure AI-generated content, and the gap is wider in B2B verticals with high expertise barriers.
And generic keyword-targeted posts are dying. Not slowly. Fast. The March 2026 update accelerated a trend that was already visible: "best X for Y" and "complete guide to Z" content that doesn't add a new perspective is increasingly left out of the index entirely.
Recalculating for Your Team
If you're running a sub-$2,000/month content operation, here's the exercise we'd recommend.
Pull your Search Console data for the last 90 days. Count how many posts published in that window are currently indexed. Divide your total content spend by the number of indexed posts. That's your real cost per indexed asset. Now look at how many of those indexed posts have generated at least 10 organic clicks. Divide again. That's your cost per traffic-generating asset.
For most two-person teams we've talked to, the second number is 2-4x the first. Some teams see 5x or higher, especially if they've been running high-volume AI generation without strong editorial filters.
The uncomfortable conclusion: publishing fewer posts with higher original-value density is almost always more cost-effective than publishing more posts at a lower per-unit cost. A team producing 5 posts per month with an 80% indexing rate spends $500 per ranked asset. A team producing 15 posts per month with a 40% indexing rate spends $750. More output, worse economics.
The Metric That Should Replace Cost-Per-Draft
We've started using "indexed-asset yield" as the primary budget metric. It's simple: the percentage of published content that enters Google's index within 30 days and generates at least one organic click within 90 days.
For teams spending under $2,000/month, an indexed-asset yield below 50% means you're burning more than half your budget on content that will never generate organic return. Above 70%, your editorial bar is working. Above 80%, you're running a tight operation.
This metric doesn't care whether you're using AI tools, freelancers, or writing everything yourself. It measures the economic output of your content investment regardless of production method. And it exposes the real cost structure that cost-per-draft hides.
Where This Leaves Two-Person Teams
The math is clear, but the trade-offs are genuinely messy. Publishing fewer, higher-quality posts means slower topical coverage. It means saying no to keyword opportunities. It means spending more time on research and original data collection, which is exactly the work that time-starved teams are trying to avoid.
We do not think there's a clean answer here. AI tools are delivering 10-20% higher ROI for companies using them in marketing, and the productivity gains are real. But the gains accrue to teams that use AI to accelerate original thinking, not to replace it.
The question for Q3 2026 planning isn't "how much content can we produce?" It's "what percentage of our production budget survives indexing to generate any return?" Until you can answer that with a number, your content ROI model is built on an assumption that stopped being true in March.
