Strategy
This is our own strategy.
Below is the plan our Strategist wrote for Propaganda's own blog, the day we started it, 14 October 2026. We answered its questions over three rounds and approved the fourth version. Every word and number in it is the real plan, copied from our account.
01 · Summary
One sentence you can repeat in a meeting.
The plan opens with what we will do, for whom, and who we want to beat to them. Ours names the buyer, the searches they type and the competitor we want them to find after us.
02 · The numbers
Three numbers, each with its reason.
Open "Why" under any number and you get the reasoning and the data behind it. Ours aims for two searches on page one this quarter, because a domain with no history can win the two under difficulty 20 in eleven weeks. The other eight take longer, and the plan says so.
03 · Topics
Four topics, so readers and Google know what you're about.
Each topic gets a range of posts, a reason grounded in your answers and the search data, and an angle a writer can argue.
04 · Launch
A new blog opens with a burst.
Fifteen posts written in twenty days and published over thirty, in three batches of pitches. Google gets something to index, and the blog is never quiet for long while it learns who reads it.
05 · Searches
Ten searches worth winning.
Long-tail first, because a new blog ranks slowly. Volume, difficulty and position come from DataForSEO that week, and every search carries one line on why it made the list. We asked to keep "ai cms" even though it brings people shopping for a free blog. It stayed.
06 · Watched sites
The blogs we read every day.
Four sites whose posts our buyers read before they compare tools. When one of them publishes on our topics, we get a pitch to answer it that week.
07 · Approval
Nothing starts until you approve.
Ask for changes in your own words, as many rounds as it takes. When the Strategist can't decide something alone, it asks. Our last answer was "forget about it", and the plan went ahead without that post.
as made in PPGD itself
What we're going to do, why, and how, in a nutshell.
This quarter we open the blog with fifteen posts on why founders cancel AI writers over wrong facts, then one post a week until mid-December, so a founder searching for an AI blog writer or an AI CMS finds us before Jesper does.
Up to four, so readers and Google can tell what you're about.
8 to 10 posts
This is what a buyer reads before paying for Jesper, AirOups or us, and the posts can show real wrong prices from real tools, which is a stronger reason to read than a feature list.
A founder cancels an AI writer the first time a post states a price or a feature the company never had, long before the prose bores anyone.
4 to 6 posts
Nobody in the data writes this for a company of twenty people, and each post can show the real answer for a real company, so the series earns the LLM searches you want without a head term.
ChatGPT describes a small company from whatever pages it happened to find, so a founder who has never checked is usually being described by an old pricing page.
4 to 5 posts
Every post can open with one real fumbled answer from your own calls, names removed, which no tool blog can copy, and it owns the generator searches that make up a third of the list.
The question a prospect asks on a sales call is the search a stranger types next month, and a fumbled answer to it is a post nobody has written yet.
3 to 4 posts
The AI CMS searches you asked to keep land here, and page one for them is a government site and an enterprise CMS, so a post on a $0 CMS that flags a stale price has no competition yet.
Every post a company publishes stops being true the day a price or a product changes, and nobody goes back to fix it because nobody knows which posts said the old thing.
15 posts written in 20 days and published over 30, so the blog is never quiet for long.
Long-tail first: a new blog ranks slowly.
Volume. How many times a month people search it.
Difficulty. 0 to 100. Under 30 is winnable for a new blog; above 60 belongs to the big sites.
Position. Where you appear on Google today.
| Target search term | Volume | Difficulty | Position | Why this one |
|---|---|---|---|---|
ai blog writer | 320 | 27 | Not in the top 100 | 320 a month at difficulty 27, the people comparing Jesper and AirOups before paying; a post on why the first wrong price gets a tool cancelled meets them mid-decision. |
ai blog write | 320 | 25 | Not in the top 100 | The same buyer typed shorter, 320 a month at difficulty 25; it is the closest winnable form of 'AI blog', which at difficulty 41 goes to Amazon Science and a Reddit thread today. |
best ai blog writer | 70 | 27 | Not in the top 100 | 70 a month at difficulty 27, a comparison search; a post on what Jesper, AirOups and Propaganda each do the day your price changes answers it by name. |
best marketing software for small business | 70 | 15 | Not in the top 100 | 70 a month at difficulty 15, the one search in the data where the buyer is a small company rather than a blogger, and a post on what a company of twenty needs from marketing software answers it. |
how to get your business mentioned in chatgpt | – | – | Not in the top 100 | Your 'get seen in LLM results' has no entry in the data, so this is our phrasing of the question a founder types; it gets looked up before this proposal is checked, and a post that shows what ChatGPT says about three real small companies answers it. |
ai blog post generator | 140 | 11 | Not in the top 100 | 140 a month at difficulty 11, the cheapest search in the data; a post on a generator that starts from your sales calls instead of a prompt box answers it. |
ai blog generator | 170 | 19 | Not in the top 100 | 170 a month at difficulty 19, the same searcher; a second post on how a generator picks what to write from one call transcript covers it. |
automate content marketing | 170 | 18 | Not in the top 100 | Your own search, 170 a month at difficulty 18; StoryChafe and SemRanch hold page one with tool lists, and a post on what a founder can automate from recorded calls is an answer none of them give. |
ai cms | 140 | 17 | Not in the top 100 | 140 a month at difficulty 17, kept at your word; page one today is ai.cms.gov and Kontint.ai, so a post on a $0 CMS where the AI flags your posts when a price changes has an open slot, and telling Claude to run a WordPress site is the comparison that post makes. |
cms ai | 170 | 18 | Not in the top 100 | 170 a month at difficulty 18, the reversed form of the same search; the AI CMS post plus a second one on what a CMS should do when a fact changes covers both. |
Goals
Then we hold you to it.
An approved plan becomes four goals, scored every day for the rest of the quarter. Our own blog is a few weeks old, so this part shows a sample company, Fullhouse, ten days before the end of a quarter.
Volume
Are we publishing enough, on the right topics?
One number for the quarter, split by topic. The line climbs with every post and the dotted line is the pace you need. Each bar shows what's done in black, what's pitched in grey and the low end of the goal as a tick. We never propose more than 20% over what you shipped last quarter.
Consistency
Is what we publish true and aligned?
A letter grade for your content, checked against your knowledge base, and one for the knowledge base itself. A is 95% clean, B 85%, C 70%, D 50%. Nobody gets 100%, and a problem you mark "not worth fixing" still counts.
Readership
Is anyone paying attention?
Total minutes read across your blog is the number we hold you to, with pageviews, pages a session and time per post beside it. Time read is the honest one. A new blog gets its target four weeks after the first readers arrive, from real numbers.
Ranking
Do we show up where it counts?
Your ten searches, checked every day. How many are on page one of Google, and for the same ten, whether Google's AI Overviews and ChatGPT (US English) mention you.
History
Tracked daily. No fudging.
Change a goal mid-quarter and the whole quarter is recalculated with the new targets. The day you moved the goalposts is marked on every graph, with who approved it and why.
Are we publishing enough, on the right topics?
Is what we publish true and aligned?
Is anyone paying attention?
- Pageviews
- 9,412
- Pages a session
- 1.84
- Time per post
- 2m 25s
Do we show up where it counts?
- staff rota template#4
- shift swap app#9
- overtime rules for restaurants#7
- restaurant scheduling software#14
- employee scheduling app#22
- clinic staff rota#31
- break rules for under 18snot ranked
- labour cost percentage#48
- how to reduce staff no shows#12
- payroll integration for rotasnot ranked
Every change recalculates the whole quarter with the new targets. The graphs mark the day.
- Version 2 · 31 Oct · approved by Maya OkaforIntegrations: 3 to 5 posts became 2 to 4Labour costs: 5 to 6 posts became 6 to 7"The POS integration slipped to January. Put the posts into labour costs."
- Version 1 · 1 Oct · approved by Maya OkaforFirst goals for the quarter.