Maña Maps passes 300 external sessions in October, up 16.7%

October closes with 314 external sessions against 269 in September: 16.7% more. That averages 10.1 sessions a day, with a median of 7. The daily series —summarised by week in the chart— has two peaks of 27 sessions, on 16 and 26 October, and a final week of 82 against 57 in the first full week. The 31 October figure is partial.

What the numbers say

External growth is real, but total sessions are no gauge. October logged 1,179 sessions, of which 865 were internal or quality-control traffic: 73.4%. In September those sessions numbered 17, so the jump reflects instrumentation and testing, not demand. Mixing both series would suggest growth that does not exist.

Read it two ways. External sessions grow at a double-digit rate from a small base, and data quality improves: the system now separates test traffic by city and local referrer. That separation is the condition for month-on-month comparison; without it, the report would be noise.

Where the traffic comes from

Of 314 external sessions, 278 (88.5%) carry referrer information. Direct accounts for 160, or 57.6% of attributable sessions. Search engines follow with 56 (20.1%), internal navigation with 32 (11.5%) and social media with 30 (10.8%). The remaining 36 sessions arrive without a referrer and are not assigned to any channel.

Bing supplies 38 of the 56 search sessions, DuckDuckGo 8 and Yahoo 9. On social, Reddit accounts for 24 of 30, with Facebook and Instagram far behind. That concentration does not justify a multichannel plan: two signals —Bing and Reddit— deserve a bounded test.

Direct, with 160 sessions, is the largest and least interpretable category: it may include bookmarks, links without a referrer, apps and unfiltered test traffic. It is used to size the audience, not to make decisions.

Markets and catalogue

Geolocation covers 218 external sessions, or 69.4%. Mexico (49), Spain (47) and the United States (45) are effectively tied and account for 64.7% of located traffic. Germany (9) and Chile (8) trail far behind. Spanish is the declared language of 72.3% of sessions. With 96 sessions lacking a country, any market conclusion would be premature.

The catalogue grew strongly: 52 published maps, 6,934 geometries, a median of 177.5 per map and a maximum of 206. This month alone added 34 maps, 65% of the catalogue. No published map lacks a source or date, and only one does not declare its legend key.

Distribution has not kept pace: 436 accumulated map views —a historical counter, not a monthly figure—, 8 maps with zero views. The most viewed are tourism (82), Gini (52) and Barcelona libraries (41). Demand clusters around a few themes; the rest need internal links and better landings.

Product and News

The viewer shows new signals: 120 map views, 17 shares, 2 sign-ups and 3 profiles. The assistant falls from 151 messages to 66, and polygon drawing from 17 to 11. Entry to the map works; features that demand more steps do not.

News generated 96 sessions, 13 of them external: 4.1% of external traffic. The jet-fuel article led with 9 sessions, followed by Delhi (6) and petrol (4). News is a channel hypothesis, not an established channel; the test is external sessions per piece and clicks through to the map.

What comes next

November has three measurable priorities. The first is distribution before catalogue: pause new maps until 15 November and work on landings, OG cards and internal links for the 8 maps with zero views. The target is zero maps without views and at least 600 accumulated views, from 436.

The second is search: optimise the 10 landings with most impressions and link them from News and related cards. The goal is to move from 56 to 100 monthly search sessions and cut Bing’s share below 70%. The third is to activate the map→account funnel, with a save CTA in the viewer and one use case per piece: 6 sign-ups and 30 shares by 30 November.

There is one decision not to take: no additional social channel and no public KPI based on total sessions, because it mixes tests with real traffic. The next review is on 30 November, with the same methodology and the external series as the reference.