How to measure if your Pump.fun volume bot actually worked

The volume number is the one metric guaranteed to look good, because you bought it. Judging a campaign by it is circular: you spent money to produce a figure and the figure appeared, which tells you the transactions executed and nothing else. The signals that answer the real question all concern what other people did while your campaign ran, and most of them can be read directly from the chain.

Volion Research Updated Jul 30, 2026 6 sections

The metrics that prove nothing

Total volume, transaction count and unique wallet count are all outputs of your own configuration. You chose the trade sizes, the frequency and the fleet size, so these figures were determined before the campaign started. They confirm the campaign ran; they say nothing about whether it achieved anything.

This matters because these are exactly the numbers most campaign reports lead with, and they are impressive by construction. A report showing large volume, high transaction counts and many participating wallets is describing the invoice, not the outcome.

Chart appearance falls into the same category. A chart looks busy because you made it busy. It is the intended mechanism rather than evidence that the mechanism led anywhere.

One qualification: these figures are worth checking for a different purpose, which is verifying you received what you paid for. Cost per landed swap, in particular, is worth computing from the actual results rather than from the quote, and it is covered in the cost breakdown.

The one signal that actually matters

Transactions from wallets that are not part of your fleet. This is the only number that reflects a decision someone else made, and it is therefore the only evidence that the activity you bought produced anything. Everything else in a campaign report is your own money reflected back at you.

Measuring it is straightforward. You know your fleet addresses. Every trade on the pair that did not come from one of them was made by someone else. Count those, and compare the rate before, during and after the campaign.

Three patterns and what each means:

Organic trades rising during the campaign. The activity is being noticed and acted on. This is the result you were paying for.
Organic trades flat throughout. The campaign produced visibility that nobody acted on. Continuing to spend will not change this, and the sooner it is recognised the less it costs.
Organic trades rising and then stopping when the campaign does. The activity was drawing other automated participants rather than people, which is common and produces no lasting effect.

Holder count, and how to read it properly

Net new holders who are still holding some time after the campaign ended is a stronger signal than holder count during it. Wallets that bought during a campaign and sold within the hour were trading the activity, not joining the token, and counting them as a result overstates what happened.

The distinction matters because raw holder count during a campaign rises for reasons unconnected to any lasting interest. Short-term traders, other bots and momentary curiosity all register identically to genuine adoption while they last.

The version worth measuring is a holder count taken some days after activity stopped, compared to before it started. That figure has had time to shed everything transient, and what remains is what the campaign actually added.

It is also the harshest measurement available, which is why it is rarely in a campaign report.

Placement and impressions

Whether the token appeared on the activity-sorted surfaces it was invisible on before is a legitimate intermediate result. It is not the final outcome, but it confirms the mechanism worked: the campaign bought attention, which was the transaction it was meant to perform.

Treat this as a checkpoint rather than a conclusion. Placement achieved with no organic follow-through means the campaign did its job and the token did not convert, which is useful information about the token rather than about the campaign.

What determines placement on each surface is covered in how DEXScreener trending works and the Pump.fun trending algorithm. Which metric each one weights determines whether transaction count or notional volume was the thing worth optimising, and that varies.

When to judge the result

Check the organic transaction rate within the first few hours, because that is early enough to stop if the answer is clear. Judge holders and lasting effect several days after activity ends. Judging anything at the moment a campaign finishes measures the campaign itself rather than what it left behind.

The early check exists to prevent the most expensive failure mode, which is continuing to spend after the answer is already visible. If several hours of sustained activity has produced no organic trades at all, more hours will not produce different information.

The late check exists because everything looks successful on the final day. Activity is at its peak, the chart is at its busiest, and none of it has been tested by the absence of spending. A few days later it has.

What the chart typically does in that interval, and how to read the decline honestly, is in after a campaign ends.

Reading a campaign report critically

Check whether the report distinguishes transactions sent from swaps that landed, since those differ substantially on high-failure venues and only the second produced anything. Check whether it separates your own activity from everyone else's. A report that does neither is describing the invoice.

Three questions worth asking of any report, including one you generate yourself:

How many swaps landed, not how many were sent? On a contested venue the gap between these is large, and the fee was paid on both.
What was the cost per landed swap? Total spend divided by successful swaps. This is the figure that connects a budget to a result.
How much of the reported activity was mine? If the answer is all of it, the report has measured the purchase rather than the outcome.

We publish measured per-venue failure rates precisely so the first two questions have a reference point that is not supplied by the party being paid. The data is on the measured transaction costs page, and the reasoning behind publishing it is on the page for our Pump.fun volume bot.

Before running anything, the calculator gives the expected cost per landed swap at current rates, which is the figure to compare the eventual result against. Campaigns are configured in the dashboard, and the honest limits of what any of this achieves are in does a volume bot actually work.

Questions

How do I know if my volume bot worked?

Count transactions from wallets that are not part of your fleet, before, during and after the campaign. That is the only figure reflecting a decision someone else made. Volume, transaction count and wallet count are all outputs of your own configuration and prove only that the campaign ran.

Why is the volume number not a measure of success?

Because you paid for it. The figure was determined by the trade sizes and frequency you chose before anything started. It confirms the transactions executed; it says nothing about whether anyone noticed.

When should I check whether a campaign is working?

Within the first few hours for organic transaction rate, which is early enough to stop if nothing is happening. Then several days after activity ends for holders and lasting effect, since everything looks successful on the final day of spending.

What holder count should I measure?

The one taken several days after activity stopped, compared with before it started. Wallets that bought during the campaign and sold within the hour were trading the activity rather than joining the token, and counting them overstates the result.

What should a campaign report contain?

Swaps that landed rather than transactions sent, cost per landed swap, and a clear separation between your own activity and everyone else's. A report missing those is describing what you bought rather than what it achieved.

What if organic activity never appears?

Stop. If several hours of sustained trading has produced no transactions from outside your fleet, additional hours will not produce different information. A campaign amplifies existing interest and cannot manufacture it.

Does a volume bot actually work After a campaign ends Full cost breakdown How DEXScreener trending works Makers versus volume Measured transaction costs