The Fallacy of Average Position in Local Pack Analytics

If you have spent more than a year calibrating local SEO campaigns, you have likely encountered the seductive simplicity of the average position metric. It appears in Google Search Console, in every rank-tracking dashboard, and in the mouth of every client who wants a single number to prove progress. Yet for anyone serious about measuring local pack and map ranking performance, average position is not just unhelpful—it is actively misleading. The local pack is not a linear list. It is a dynamic, personalized, proximity-weighted surface that shifts based on device type, time of day, query refinement, and even whether the user has previously clicked a given listing. Reducing this multidimensional reality to a single decimal point is an act of analytical violence that obscures the signal you actually need.

Consider the geometry of the local pack. On mobile, the 3-pack occupies the top third of the screen, but the gap between position one and position three in terms of visibility is not an arithmetic progression. Position one often receives more than forty percent of clicks, while position three hovers in the low teens. But that distribution is not static. If a searcher is physically inside the service area of a competing business, Google may bump that competitor into the first slot regardless of your authority signals. That means your average position can fluctuate wildly based on a variable you cannot control—proximity. A rank-tracking tool that queries from a fixed location in downtown Chicago will report position two for your client in the suburbs, while a user actually standing in those suburbs might see your client at position one. Average position becomes a phantom statistic that correlates only weakly with actual user experience.

The deeper issue is that Google does not rank local pack results on a single axis. The algorithm blends traditional organic factors like relevance and prominence with a decay function tied to distance. This means that even if you dominate every citation and review signal, a user who is twenty miles outside your service radius may never see you in the pack at all. Your average position across all queries then gets pulled downward by these out-of-range impressions, creating the illusion of poor performance when in reality your conversion rate from in-area searchers is excellent. The takeaway is simple: raw position data must be segmented by query-to-business distance before it holds any interpretive value.

Furthermore, the local pack itself is not a stable entity. Google often replaces it with the local finder—a full map overlay with its own UX hierarchy—when the query is ambiguous or when the user zooms in. In the finder, positions are dynamic and scrollable, and the concept of a fixed rank dissolves entirely. Measuring position in a finder context is like measuring the wave height of the ocean with a ruler you hold at an angle. The metric breaks. Savvy analysts have started tracking impression share within the pack rather than ordinal rank, using data from the Google Business Profile Insights panel combined with third-party tools that capture whether a listing appears at all for a given search. Impression share tells you what percentage of relevant searches your business is visible in, which directly correlates with opportunity. Position then becomes a secondary diagnostic, not a primary KPI.

Another layer of complexity is personalization. Google customizes local results based on search history, device data, and even patterns of previous interaction with specific businesses. Two users standing at the same street corner can see different pack orders for the same query. Your rank-tracking tool, which likely uses a single headless browser session with no search history, reports position one. But the actual local audience on mobile devices with varied histories may see you at position three or not at all. The solution is not to abandon tracking but to triangulate: compare your tool’s results with Google Business Profile’s own performance reports, run manual queries from multiple devices, and build a regression model that accounts for historical click data at the query level.

Finally, we must discuss the temporal dimension. Local pack rankings shift throughout the day based on business hours, real-time location signals, and freshness of reviews. A dentist’s office that updates their hours for a public holiday may see a temporary surge in pack position for that day, only to revert the next morning. Averaging this over a week gives you a number that doesn’t correspond to any real user experience at any specific moment. Instead, consider using a time-decay weighted visibility score that prioritizes peak conversion hours. If most of your customers search between 9 AM and 5 PM, measure pack presence during those windows exclusively.

The path forward requires an upgrade in your measurement philosophy. Stop asking “What was my average rank?“ and start asking “For what proportion of high-intent, in-area queries did my listing appear in the pack, and what was the distribution of its vertical displacement relative to the nearest competing business?“ This is a tougher question to answer, but it is the only one that reflects the actual competitive dynamics of local search. When you present this nuanced data to stakeholders, frame it not as complexity for its own sake but as precision—the difference between a weather forecast that says “some rain possible” and one that tells you exactly when and where to bring an umbrella. Your local pack performance is not a rank. It is a probability surface. Measure it accordingly.