User:BrownHairedGirl/Cullportals

Draft RFC

Should the system of portals be reduced to only the 11 portals linked from the Wikipedia Main page?

The portals to be kept are:

three portals on the left-hand column of the main page: Portal:Current events, Portal:Contents, Portal:Featured content
eight portals linked on the top right of the main page: Portal:Biography, Portal:History, Portal:Arts, Portal:Technology, Portal:Mathematics, Portal:Science, Portal:Geography, Portal:Society

The remaining ~870 portals (see Category:All portals) would be marked as historical, and would cease to be linked to from articles, categories, and the 11 remaining live portals.

Rationale

Because only these 11 portals which meet both

the purpose set out in WP:PORTAL, that "Portals serve as enhanced 'Main Pages' for specific broad subjects"
and
the scope criterion set out in WP:POG, viz that portals should be about "broad subject areas, which are likely to attract large numbers of interested readers and portal maintainers".

Table 1. Top 15 portal pageviews April–June 2019 Source: massviews
Rank	Portal	Daily average	Main page?
1	P:Current events	48,023	Yes
2	P:Contents	15,804	Yes
3	P:Featured content	6,267	Yes
4	P:Biography	2,473	Yes
5	P:History	2,465	Yes
6	P:Arts	2,069	Yes
7	P:Technology	2,000	Yes
8	P:Mathematics	1,849	Yes
9	P:Science	1,760	Yes
10	P:Geography	1,620	Yes
11	P:Society	1,235	Yes
12	P:Companies	332	No
13	P:LGBT	311	No
14	P:Food	304	No
15	P:Aviation	229	No

Table 2. Distribution of pageviews April–June 2019 Source: massviews
Daily average pageviews	Number of portals	% of portals
> 1000	11	1.22%
501–1000	0	0%
251–500	3	0.33%
101–250	42	4.64%
51–100	90	9.96%
26–50	181	20.0%
<25	577	63.8%

The numbers are simple. The Table 1 shows the daily average viewing figures for the top 15 portals. The first 11 portals are all linked from the front page, and all get over 1,000 page views per day.

But the 12th-ranked portal is Portal:Companies, which is not linked from the front page, and it gets only a quarter of the views of the lowest-ranked front page portals.

Thereafter, the numbers tail off rapidly. See Table 2

94% of portals receive less than 100 pageviews per day.

64% of portals receive less 25 pageviews per day.

Not malleable

These figures will not change significantly.

In April–June 2019, the main page got an average of 16,383,063 views per day. There were 85,565 views of the 11 portals linked from the front page. So roughly one in every 200 mainpage readers chose to visit a portal.

That one-in-200 ratio is very similar to the ratios repeatedly seen in research done for MFD discussions over the last 5 months. With the portals examined there, the ratio of portals views to head article views has overwhelmingly been between ~1:100 and ~1:2000. (That's from memory. Check, and ask with @Robert McClenon's logs)

These numbers are not significantly malleable. Only a very few exceptional portals have broken significantly out of that range.

That means that the overwhelming majority of portals receive very low page views, and will continue to receive very low page views, because the topics which they cover are too narrow. Some portals which are linked to from very high numbers of pages may receive higher than average page views, but in general only very broad topics with exceptional promotion break out of the cycle of low page views. And even those portals still come nowhere near the viewing level of the front page portals.

Vicious cycle

Portals require ongoing maintenance. Articles are moved, assessments change, topics need updating. The "mini main page" model of portal is based on the Wikipedia main page. But the WP main page is the product of huge amounts of ongoing work by several large and busy teams of editors, at WP:DYK, WP:TFA, WP:TFP etc. Their work is thorough, and it serves a huge audience.

But most portals get tiny audiences. Of the portals which still remain in July, the median portal for pageviews in April-June 2019 was Portal:Houston, about the city in Texas. It got 17.67 pageviews per day. 440 portals get more than that; 440 gets less than that.

Special:PrefixIndex/Portal:Houston shows a set of subpages mostly dating from a monthly update model abandoned a decade ago. The topics actually displayed are chosen from a set of 11 "selected articles". The most recent of those was in Portal:Houston/Selected article/11, which was added in January 2013, and remains unchanged since 2013.

There is also a set of ten selected biographies. I looked only he first one: Portal:Houston/Selected biography/1, about Yao Ming. It says he plays for the Houston Rockets ... but actually he left that team in 2011.

Some editors have worked on the presentation, so it looks polished. But under the cover, it's basically a very dusty old magazine archive: a set of content forks created between 2011 and 2013, laid out above a pile of abandoned stuff from even earlier.

That's just a little over one millionth of the views of the Wikipedia main page. So it's no wonder that this portal is languishing, as are so many others. Why put effort into a showcase which almost nobody reads?

Houston is not a cherrypicked example. It's the precise midpoint of the set of remaining portals.

In the last 5 months, a series of MFDs deleted 4,200 automated spam portals created in 2018. Over 600 more pre-automation portals have been deleted at MFD since then, because they have been abandoned, in some cases for up to 14 years.

So we now have just under 900 portals, compared with 5,705 at the end of the portalspamming era. After weeding out the worst, we are down to the last 16% of portals ... and the very midpoint of the remainder is an abandoned portal with under 18 readers per day, who are being served a diet of decade-old content forks.

This nobody's fault. There are simply far too many portals with far too few readers, which in turn means far too little maintenance