For most research, testing the market is routine. When ad-hoc studies get bid, sample suppliers get compared and the work goes to wherever fits the project in front of you . Trackers are the exception. Once a tracker is running, its sample supplier tends to carry forward for the life of the tracker.
That is worth reconsidering, because the conditions that made a supplier the right choice at launch do not stay fixed. Technology changes, leadership changes, priorities change, and the market for sample shifts with them, supplier by supplier and across the industry. A partner that fit the program on wave one may not be the best fit for the next wave, and there is no way to know without looking.
Two things keep most teams from looking. Re-bidding a live tracker feels prohibitively difficult and time-consuming. And continuity is the deliverable, the series is the asset, so a visible step in the trend outweighs most of what a change could offer. Both concerns are real. Neither is as prohibitive as it seems.
Bringing on a new sample partner is worth doing carefully, and it can be done without disrupting a live program. The goal is straightforward: confirm a new source lines up with what you already have, before anything moves. A good supplier will have a defined process for doing it.
It happens at scale. A leading experience management company moved more than 100 trackers to a new sample blend across 20+ countries in four months, with brand funnel metrics coming back more consistent than historical data.
How to evaluate changing providers
There are two clean ways to evaluate a new provider without disrupting your program.
- Run a side-by-side. Field the new sample provider alongside your current one on a wave and compare the key metrics directly. This is the most direct test for an existing tracker moving from one supplier to another.
- Introduce sources gradually. Phase new sources into the blend over time, confirming the read holds at each step. This works better for programs where a full parallel test is not practical.
Either way, the decision rests on data. And if the budget is already set for the year, the timing can align with your renewal: the comparison runs now, ready when the window opens.
A rigorous parallel test starts with your current sourcing rather than with a proposal. Before anything fields, there should also be agreement on what success looks like and on the path to transitioning the program if the results hold. Without that, a parallel test is an exercise with no defined next step. The guide covers what to expect at each stage, and what to ask for.
What renewal leaves unexamined
A parallel test tells you how a new sample provider compares against your current one. It does not diagnose what has been happening inside the current one, and on most programs you could not check that anyway, since the composition is not disclosed. The four points below are reasons to run the comparison rather than findings it produces, and each one accumulates quietly in programs that carry forward year to year.
- Undisclosed blending. Most programs already run on blended sample, disclosed or not. When a provider adjusts the mix to fill quota, the baseline moves underneath the trend, and a shift you cannot see is a shift you cannot diagnose.
- Respondent overlap. Because respondents belong to more than one panel, the same person can enter through different sources. Left unmanaged across waves, that overlap erodes the independence each reading depends on.
- Quality drift. What a screening approach needs to catch keeps changing, from new fraud patterns to AI-written open ends, so quality has to be maintained continuously rather than set once. On a tracker the standard has to hold in wave twenty exactly as it did in wave one, or quality decay reads as real change. PureScore™ is respondent-level rather than session-level, applying the same standard to every respondent, on every wave, in every market, and updated as new threats appear.
- Pricing inertia. Trackers rarely get re-bid, so renewal pricing drifts above market by default, and a rate carried forward without contest for several cycles is one nobody has tested recently.
On whether a blend holds steadier
The concern underneath all of this is usually whether multi-sourced sample is less stable than a single panel. Independent three-wave research found the opposite. Individual panels shifted materially between waves on their own, while a deliberately managed blend cancelled those shifts out. The findings are in the guide.
Five elements to review in a program you already run
Whether you end up moving or staying, the review itself comes down to five things, and they matter in combination rather than individually.
- Diversified supply. A narrow source mix means shifts in one source show up as changes in your trend.
- Source transparency. Without source visibility, you cannot explain a wave that moves unexpectedly.
- Service and delivery. How a supplier handles a difficult wave tells you more than how they handle an easy one.
- Modern data quality. Every provider screens; on a tracker, what matters is the approach — quality checked at the respondent level rather than the session, and applied the same way on every wave.
- Competitive pricing. Tracker rates hold steady by default, not because they are still right.
None of these hold up alone. A diversified supply base only helps if you can see the sources in it, and that visibility only matters if someone acts on what it shows. Screening and pricing follow the same logic: neither means much without the delivery to back it.
A tracker is only as good as its sample
The continuity that makes a long-running tracker valuable is what makes it feel unavailable to change. The constraint is narrower than it looks. With a defined comparison process and full source visibility before and during it, what you want to know can be established before anything in the program moves. None of that requires a decision to move. It requires a decision to look.
The working guide walks through each of the five, with the questions to ask and what a solid answer looks like, and covers how to evaluate a new source against a live program.