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

Brand Marketing › Advanced

Checkpoint: Brand Investment and Long-Term Effects

Scenario questions set in an invented brewery test whether you can defend brand spend with evidence rather than belief. They cover budget drift towards performance, short-term and long-term effects in the data, the blind spots of click-based attribution, adstock and saturation, the history a mix model needs and how to combine several measures into one argument for a board.
CheckpointAdvanced levelModule 1: Brand investment and long-term effects4 minUpdated

What this checkpoint covers

Bellfounder Ales is an invented independent brewery. It sells through supermarkets, pubs and its own online shop, sponsors a summer folk festival, runs online video in the spring and buys paid search and retailer promotions all year. Its finance director wants every pound defended. The questions draw on the brand versus performance guide, the marketing mix modelling basics guide and the comparison of mix modelling and multi-touch attribution.

Task

Write a short answer and a reason for each question before opening the explanations. Count a right answer with a wrong reason as a miss, and reread the linked guide for every miss.

Questions

  1. Over two years Bellfounder moved money from the festival and video into paid search and deeper retailer promotions, because those lines report the best returns. Volume is holding. Does that prove the move was right, and what would you check?
  2. After the festival, online orders that month were flat, but branded search and requests from new pub accounts rose slowly over the following months. How should each kind of effect be read?
  3. The data-driven Attribution report gives the festival and the video almost no credit and gives brand search a great deal. Why, and where has the missing credit gone?
  4. A mix model finds a long carry-over for video and an almost flat response curve for paid search at current spend. What does each finding mean for next year's brand budget?
  5. Bellfounder has a year and a half of monthly sales, and video spend was the same in every spring. An agency offers to build a mix model now. What do you advise?
  6. The finance director proposes switching all brand activity off for a quarter to see what happens. Is that a good test?
  7. The board wants one argument for restoring brand spend. How do you combine the measures available into it?

Answers explained

  1. No; holding volume on deeper discounts is a warning sign. Budgets drift towards capture because reports flatter it, not because it is proven. Check whether cost per sale in capture channels is rising, branded search is flat or falling, the share of new buyers is shrinking and volume now depends on promotions. See signs the balance has tipped.
  2. The festival was not an activation. Short-term effects show as orders in the weeks of spend and stop when spend stops; brand effects arrive as slower movement in branded search, new accounts and later sales costs. Judge the festival on lagged measures over quarters, not on the month's orders. See timescales of effect.
  3. Neither leaves a click. Click-based models only share credit among recorded touches. People who saw the video or the festival and later searched the brand appear as brand-search conversions, so the credit lands on the last recorded touch. See offline and brand effects.
  4. Video works over weeks; search is saturated. Adstock means video's effect outlasts its weeks on air, so short reporting windows under-credit it. A flat curve means the next pound in search buys little, whatever its average return. Shift at the margin and plan on marginal return. See adstock and saturation.
  5. Not yet. A model needs a consistent, usually weekly, history spanning more than one seasonal cycle, and variation in spend to learn from. Constant spring video coincides with the season, so the model cannot separate them. Plan deliberate variation, weekly data and controls for price, promotions and weather first. See data requirements and limits.
  6. Not in that form. A national switch-off has no comparison group, mixes the effect with season and competitor activity, and risks real loss for a muddy answer. Vary brand activity by region instead, with comparable regions held at the old level, and read the result over a long enough window for slow effects. See using both with incrementality tests.
  7. Triangulate. Brand tracking shows whether memory moved among category buyers; branded search shows whether it turned into behaviour; capture-channel costs show where the return is cashed; a regional test gives a causal read; and once the history allows, a model (MMM) calibrated to that test sets the channel split. Add Share of Voice against share of market for the competitive context. See measuring brand honestly.

Where to go next

If the diagnosis questions caught you out, redo the brand investment exercise for a mortgage broker. Then move to the next module with the cross-platform reach exercise for a cider brand. To revisit awareness measurement and trackers, return to the intermediate brand marketing level.

Pages this lesson applies

Terms used on this page

Also related

Related