SaaS · 6 months

Content automation: four articles a month from one person instead of three

B2B service, unnamed

  • AUTO
  • STRATEGY
4+100%
2Feb
3Mar
3Apr
4May
4Jun
4Jul
Початок: 2Кінець: 4
Крайні точки: реальні дані з проєкту. Проміжні місяці відновлені за кривою зростання, щоб показати динаміку, і не є окремими замірами.

About the project

A B2B service running its blog with three people: a marketer found topics, a writer wrote, a designer made the illustrations. The output was two articles a month, and all three believed they were at capacity.

The challenge

The client asked us to hire another writer. We asked for a week to measure the process before hiring anybody.

What the measurements showed

We broke the production of one article into stages and timed them:

Stage Hours Who
Finding a topic and checking demand 3.5 marketer
Gathering sources and facts 4 writer
Writing 6 writer
Illustrations 3 designer
Layout and publishing 2 marketer
Internal linking 1.5 marketer
Social announcements 1 marketer

Twenty-one hours. Six of them on the actual writing.

So 71% of the time went to everything except the text. Hiring a second writer meant widening a bottleneck that did not exist.

What we did

Finding topics: from 3.5 hours to 20 minutes. We built a keyword map six months ahead once and set up a monthly position check. Topics are no longer searched for, they are taken off a list in priority order.

Gathering sources: from 4 hours to 1. We built a shared base of verified sources and industry figures. Instead of searching again, the writer takes from it and adds what is new.

Illustrations: from 3 hours to 15 minutes. We built chart and diagram templates in one style. The designer no longer draws every image; once a quarter they refresh the set.

Publishing and linking: from 3.5 hours to 40 minutes. An article template with ready blocks and automatic insertion of links to related material by tag.

Announcements: 1 hour to 0. Publishing an article now creates draft posts for three channels. A person only proofreads them.

What worked

The biggest effect came not from tools but from measuring first. Three people spent a year certain that writing was the bottleneck, because writing is what feels like work. The measurements showed writing took under a third of the time.

What went wrong

Quality dipped for the first two months. Handed a ready topic and ready sources, the writer started writing faster and shallower: the immersion that used to happen while hunting for material disappeared.

We had to put one stage back: the writer personally verifies two key facts in every article, even if they are already in the base.

What we should have done

Automate the routine and leave alone the things that look like routine but are actually part of thinking. Finding sources is routine. Checking that a number is real is not.

What a proper plan does to the economics

Before: 2 articles × 21 hours = 42 hours a month, three people. After: 4 articles × 7 hours = 28 hours a month, one person.

That freed roughly 1.3 full-time positions while doubling output. Organic traffic grew from 1,400 to 5,900 a month over six months, not because the articles got better, but because there were twice as many and each one shipped on time.

Get a consultation