Measuring Marketing Effectiveness for a Charitable Organization

The challenge

A charity’s income does not arrive from one source. Some gifts trace back to a specific activity. Many do not, because donors respond to a mix of paid, social and owned channels over time. A large U.K. charity needed to know which marketing activities were bringing in income, including income no tracking system could attribute, and asked MASS Analytics to measure it.

Why attributed income tells only half the story

Attributed income is the part a tracking system can link to an activity. Unattributed income is everything else, such as gifts that follow several messages across several channels. Attributed income alone rewards the easiest activities to track and undervalues those that build awareness first. The charity needed both measured together.

Four measurement gaps the charity needed to close

• Attributed and unattributed income. Fundraising affects both, and they respond differently. The charity needed the impact measured on each.
• Direct and indirect tactics. Direct fundraising and indirect marketing work together. Each needed its own measured contribution to income.
• Long-term impact. Some activities pay back beyond the financial year they ran in. The charity needed to see those effects.
• New donors versus renewals. What wins a first-time donor is not always what keeps one. The charity needed to separate the two.

Those four gaps became six measurable questions, each answered by its own model (Table 1).

The solution

MASS Analytics built six models on three years of weekly data. Income was the main measure. The drivers were marketing activities, paid media, digital and social media, owned media, and external factors and seasonality. Each model answered one of the charity’s questions, so the full picture of direct and indirect impact came from one consistent approach.

Six models for six questions

Four models measured the impact of direct and indirect activities on unattributed income, attributed income, new donors and renewals. A fifth measured the effect of marketing activities on measures that are not income, such as relevance. A sixth identified the main drivers of non-financial engagement, such as social media fans. Processed and artificial variables were added to correct for underfitting and for variables that showed no significance, so each model kept its explanatory power.

Data foundation and regional models

Data from multiple sources and structures was standardized automatically, which sped up data preparation. Where regional data was available, regional models were built as well. They add variation and data points, which raises the detail and confidence of every measure. They also capture activity that ran in one region, which a national model can miss.

Results and Impact

The models measured the return on each media channel and identified the best performers. They also gave the charity a way to plan its budget so that income can rise at the same media spend.

Return measured for every media channel

The models mapped the contribution of each factor to income and measured the direct and indirect impact of marketing and media activity. That let the charity measure the ROI of each media channel and see which ones performed best. The work also covered the cross effect between offline and digital channels, so the charity could see how they worked together.

Budget planning at the same media spend

The charity received a full model for planning how to allocate its budget, designed to raise income without raising media spend. A scenario and simulation tool came with it, so the team could test different budget allocations itself before committing spend.