Science Mission Impact: can smaller, cheaper missions do the science of big ones?

As historic proposed cuts lead to delays and mission cancellations, NASA is again looking to smaller and cheaper projects to replace them.

Could that work?

We asked four questions:

  1. Does scientific impact per dollar vary with mission cost?
  2. Do lower-cost missions fail more often?
  3. How does time to science change with cost?
  4. What is the scientific productivity of commercial lunar landers?

To answer these questions, we studied the publication outputs of 143 NASA-led science missions launched from 1980 to 2024, across all five of the agency’s science divisions.

We assembled a set of 148,744 peer-reviewed papers associated with these science missions.

Papers published in peer-reviewed academic journals are a proxy for scientific productivity.

When one paper builds on another, it cites it. Citations are one measure of scientific impact.

Papers are not cited equally.

In planetary science, the most cited paper had 260 citations. But half accrued nine or fewer citations in their first four years.

Citations are counted through the third calendar year after a paper appears. Why?

We consider a paper to be high-impact when it has more citations than 90% of the other papers in its field in our dataset.

In planetary science, that means papers with 31 citations or more are high-impact.

The rate of high-impact papers by cost

Here, 37 planetary science missions are ordered left to right by their total cost, adjusted for inflation.

denotes a total mission failure; denotes a partial failure.

Each bar shows a mission’s average number of high-impact papers published per year over its measured publication window.

Across the four divisions shown, larger missions tend to produce more high-impact papers per year, as you would expect.

The question is what they produce per dollar.

The Biological & Physical Sciences division has too few missions, so it isn’t shown here.

Missions below $150M are considered very low cost.

Some stand out: TRACE (Heliophysics) produced high-impact papers at a rate comparable to much more expensive missions.

Most produce few or no high-impact papers.

Scientific impact per dollar

Dividing each mission’s citations by cost compares citations per $100M spent. This can be thought of as scientific impact per dollar.

The best citation return per dollar in planetary science came from Mars Global Surveyor, a $531M mission, adjusted for inflation.

If lower-cost missions consistently returned more impact per dollar, we would expect the average citations to be higher on the left side of the plot.

If average impact per dollar were similar across costs, the line would be approximately flat.

This is the actual running average of impact per dollar. In planetary science, very low-cost missions returned poor results, primarily due to mission failures.

The average peaks in the mid-cost range.

How we calculate the averages

In no science division did very low-cost missions demonstrate better impact per dollar than larger ones.

Conversely, NASA’s most expensive science missions did not consistently demonstrate an advantage, either.

Shading shows a 95% interval for the average at each cost.

The same missions, grouped by cost.

Mid-cost missions, between $150M and $1B, matched or beat the most expensive missions on impact per dollar in two of the four divisions.

Bars are averages of the missions in each group, failures counted as zero, with 95% intervals.

Mission failure rates

Very low-cost missions (the first column) have a full or partial failure rate of 32.6%.

For mid-cost missions, between $150M and $1B, it is 12.9%; above $1B, 2.9%.

Time to high-impact science

Of the very low-cost missions in our study, 72% produced no high-impact papers at all.

For those that did, the median time from project start to a first high-impact paper was not significantly shorter than for mid-cost missions.

The most expensive missions took longer to return high-impact science, likely due to longer development and transit times.

Dashed lines mark median values. Shading indicates the 95% interval.

But when measuring from the start of science operations, more expensive missions tend to result in a high-impact paper almost immediately.

Scientific productivity of CLPS missions

The four Commercial Lunar Payload Services (CLPS) missions flew too recently to compare scientific impact using citations.

Instead, we can compare their scientific productivity by counting the number of papers published since launch.

Compare CLPS output with two science-first NASA missions. These also had short mission durations and mixed experimental technology demonstration with science.

So far, CLPS landers are trailing in scientific productivity.

What we found

  • Very low-cost missions do not provide a clear advantage in science per dollar, and neither do very expensive ones.
  • High failure rates. Of missions below $150M, 33% failed. Only 9% of larger ones did. Failure rates vary across scientific divisions.
  • Not reliably faster. Most low-cost missions generate no high-impact science papers. Those that do are not significantly faster than mid-cost missions.
  • CLPS has yet to demonstrate comparable scientific productivity. Each mission has, so far, produced fewer papers than comparable NASA-led science missions.

Very low-cost missions are worth doing: They help train future mission leaders, prove out new technology, and can return good science.

But nothing in this record demonstrates they can replace the scientific impact of more complex and costly missions.

Conversely, flagship missions are worth doing, too. Citation efficiency is not the only measure of value. Larger missions tend to break new ground, provide unique data, and develop technologies that enable subsequent smaller missions.

Explore the data yourself. Scroll down to access our entire dataset by division.