Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Programmatic Advertising › Beginner

Setting Brand Safety Rules for a Children's Book Publisher

A practice exercise in designing quality controls from an objective rather than from fear. You read a week of monitoring data from an invented children's book publisher's open-auction buy, decide its safety floor, suitability tiers, site lists and viewability setting, place each control at the right level of the account, and work out what the controls cost in reach and price.
ExerciseBeginner levelModule 3: Quality controls and a first campaign5 minUpdated

Why this matters

A buyer selling picture books to parents cannot afford to appear beside unsuitable content, and cannot afford to filter so hard that the campaign reaches nobody either. Every control removes inventory, and the controls compound. This exercise applies the guide to brand safety and viewability, which sets out the controls and their trade-offs, and the guide to DV360 account structure, which explains which settings belong at advertiser level and which on each line item. The settings you write here are also what a peer reviewer checks under the paid campaign launch SOP.

The publisher and its risks

Little Lantern Books is a fictional independent publisher of picture books. Its customers are parents and grandparents; its readers are young children. It buys display in Display & Video 360 (DV360) through an agency, aiming for engaged visits to its online shop. Monitoring first, the agency ran a week of open-auction delivery with only the safety floor on and a verification tag on every impression.

Illustrative example — not a real result

Little Lantern Books: one monitoring week, open auction

Inventory groupShare of impressionsMeasured viewabilityAverage CPMNotes from the verification report
Trusted national and regional news22%64%£3.10A few pages on crime and tragedy
Parenting and family sites18%61%£3.60Mostly suitable
Book, reading and library sites9%58%£3.30Fully suitable; most shop visits came from here
Parenting forums, user-generated7%49%£2.40Some threads on adult topics
Games and apps aimed at children6%72%£1.90Child-directed inventory
Unfamiliar domains with auto-refreshing units31%81%£1.20Made-for-advertising pattern; very high bounce
Celebrity gossip and mature entertainment7%55%£2.80Content rated for mature audiences

Total for the week: 500,000 impressions. Overall measured viewability: 67%. Blended CPM: £2.48.

Choosing the controls

Separate safety from suitability. The safety floor is already on and not negotiable; suitability is the publisher's own judgement above it, written down with the people who own the brand.

Then choose the shape of the buy. An inclusion list admits only named sites and is the strongest control; exclusion lists and vendor segments remove named problems from an otherwise open buy. Treat viewability as a separate question from safety: the table shows why. Finally, place each setting at the level where it belongs, so that the next line item someone creates starts safe without anyone remembering to tick a box.

Task

  1. For each inventory group, decide: keep, keep with a page-level suitability tier, or exclude. Give one reason for each.
  2. Decide whether the campaign should run on an inclusion list or on exclusions from the open auction, and say what goes on the list.
  3. Name the control that removes the made-for-advertising inventory, and explain why a viewability filter would not.
  4. Set a viewability approach suited to engaged shop visits.
  5. Say which settings belong at advertiser level and which on the line item.
  6. Work out what your rules leave of the monitoring week's impressions, the new blended CPM and the new viewability, and say what that means for the budget.

Worked answer

Group by group. Keep book and reading sites and parenting sites. Keep trusted news on an inclusion list, with the verification vendor's suitability tier excluding tragedy and crime pages, rather than blocking news by keyword. Exclude parenting forums for now: user-generated threads cannot be classified reliably enough, and they can return later with a strict page-level tier. Exclude games and apps aimed at children: the buyer is the parent, the inventory is aimed at the child, and the request's regulatory flags limit what can be used there. Exclude gossip and mature entertainment through the content rating setting. Exclude the made-for-advertising domains.

Shape of the buy. Run on an inclusion list of news, parenting and book sites. It removes the made-for-advertising long tail by never admitting it, and the verification vendor's made-for-advertising segment backs it up if the list is later widened. A viewability filter would keep that inventory, because auto-refreshing units score highest of all.

Viewability. Do not chase a high threshold. Set it near what the kept editorial groups already achieve, and optimise towards engaged visits, an outcome that made-for-advertising sites never produce.

Where settings live. Sensitive category exclusions and digital content labels go in the advertiser's brand-safety defaults, so every new line item inherits them. The inclusion list, the vendor segment and the viewability setting go on each line item, because they may differ by tactic.

Illustrative example — not a real result

What the rules leave

Kept groups: news 22% + parenting 18% + books 9% = 49% of 500,000 = 245,000 impressions a week, before the news suitability tier removes a few more.

Blended CPM of the kept groups: (0.22 × £3.10 + 0.18 × £3.60 + 0.09 × £3.30) ÷ 0.49 = £1.627 ÷ 0.49 ≈ £3.32, up from £2.48.

Viewability of the kept groups: (0.22 × 64 + 0.18 × 61 + 0.09 × 58) ÷ 0.49 ≈ 62%, down from 67%.

Budget: last week's spend was about 500,000 × £2.48 ÷ 1,000 ≈ £1,240. At £3.32 that buys about 373,000 impressions, more than the 245,000 the kept groups supplied, so the line item will under-spend unless the inclusion list grows.

The arithmetic carries the lesson. The controls cut Reach roughly in half and raise the cost per thousand (CPM) by about a third, and measured viewability falls, because the most viewable inventory was the worst. Judge the new settings on cost per engaged visit, not on raw CPM. Then widen the inclusion list with more book, library and family sites so the budget can spend on inventory that met the standard.

Common mistakes

  • Blocking news by keyword. It removes quality journalism; use a page-level suitability tier on trusted news.
  • Using viewability as a safety control. The made-for-advertising group had the best viewability in the table.
  • Setting every control on the line item. The next line item someone builds starts unprotected; put the floor and content labels in the advertiser defaults.
  • Reporting raw CPM as the result. Compare cost per engaged visit before and after.

Next, bring the whole level together in the first programmatic campaign capstone, or see the full module on the beginner programmatic level page.

Pages this lesson applies

Terms used on this page

Also related

Related