Measuring Instagram reach with barely any magic

Instagram reach can be measured. At least sort of. Here’s the SMK approach.

I’ve been describing our approach to Instagram measurement in a number of emails now, so I thought I’d just outline it here.

For those of us who like to track what we’re doing the whole problem, of course, is that Instagram does not provide any behind-the-scenes metrics. Thus, while you can add up engagement numbers (numbers of likes, numbers of comments etc.) you can’t tell how many people actually saw your masterful squares.

But that’s just one problem. At SMK, we really prefer to look beyond our own channels and include what museum guests and others are sharing. We like to see ourselves as initiators or inspirators of conversation, but we certainly don’t see worthwhile activity as confined to the channels that we happen to “own”.

Here’s what we do: First of all, we try to catch the largest possible amount of Instagram activity related to the museum. For this, we use the wonderfully flexible IFTTT (If This Then That) web service. IFTTT users can set up simple “recipes” using “ingredients” in the form of web services and selected actions. For instance, a recipe can be “Send any image uploaded to a specific Facebook page to a specific email address”.
We’ve set up IFTTT to track Instagrams posted at the two locations of SMK and/or with one of the hashtags often used by museum guests.

IFTTT recipe
IFTTT recipe

When such a photo is posted, IFTTT adds it to a Google Drive spreadsheet immediately. The spreadsheet fills up with lines like this:

Instagrams in a spreadsheet
Instagrams in a spreadsheet

At this point, what we have is just lines in a spreadsheet. Getting to reach requires a combination of math and guesswork informed assumptions.

We counted the number of followers for all posters in one whole month and it turned out they had an average of 400 followers. To measure reach for any month, we thus multiply the number of photos with 400. Obviously, however, Instagramers don’t see all images in their feed – and there’ll be a certain overlap of followers – so we’ve found it appropriate to divide by three.

Now, the ‘three’ is highly arbitrary. Is three the optimal number? Almost surely not (it is a prime number, though, shouldn’t that count for something?…). You might suspect that the resulting number will be too high, but our logic is that the IFTTT does not capture everything posted from the museum (i.e. everything not geo- or hashtagged) and so we compensate upwards. Thus, the SMK formula is:

Instagram reach = Number of photos posted in a month * 400 / 3

Of course, you can adjust the 3 any way you find appropriate.

Feel free to argue, but my logic is this: Our formula is not perfect, but it is the most reasonably approximation we can currently come up with. Also, while the actual number should be asterisked with disclaimers, the math here is simple and gives an easy way to track development over time. In short, we are happy with it until more precise tools come around.

Squaring the museum: A year of Instagram at SMK

For the past year we’ve hosted a small series of Instagram events at the National Gallery of Denmark. The result is a diversity of perspectives and quite a few masterpieces.

Given enough eyes, if you’ll pardon the paraphrase, all corners of the museum can be made photogenic. A year ago this August we had our first Instagram event at SMK, giving a group an exclusive tour an hour before opening to the public (a tour where I elegantly managed to get myself locked all alone into the permanent exhibition in ways still somewhat unexplained).

And just to be clear: It is a truly great thing. Inviting visual creativity inside is a real pleasure and a great way to engage with the art in new ways (not to mention the staircases and each and every reflective surface). In this post, I’ll sketch our experience so far.

Instagramers getting ready
Instagramers getting ready (August 2014)

For our very first attempt at the genre, Instagram staff members – in town for Copenhagen Fashion Week – actually handled logistics. A group of 12 specially invited IG’ers were taken for a “empty museum” walk under the #emptysmk hashtag. Yes, we bent our hashtag model which is always SMK-something to associate ourselves just a bit with the fabulous #emptymet series of events.

Going on instinct we’d asked a guide to lead the way, speaking knowledgeably about the collection. This effort turned out to be wasted, however, as the photography mission left very little mental room for taking in art history.

Some of the photos from this first tour are shown here:

One more from #emptysmk > discovering colour with @kjwww // #artwatchers #timeforteal

A photo posted by Hannah Waldram (@hannahrw) on

Soon after the walk we were contacted by Instagramers Denmark, and with that fine association we made a similar setup (only after hours, not before). About 15 people went for an #emptysmk walk, once again with great images to show for it:

tak @igersdenmark @smkmuseum for en hyggelig fototur i #emptysmk. A photo posted by laurenlila (@laurenlila) on

  One more from the event today // #emptysmk @smkmuseum   A photo posted by Judith Stohn (@jstohn) on


Wanting to try our hands at something slightly more focused we held our next event at the Royal Cast Collection, which is in a separate location from the rest of the museum. As it was now winter and we had to plan for late in the day, we arranged for some simple extra lighting in the form of two strong photographer’s lamps that we could move around to try different setups. Again we worked with Instagramers Denmark to invite 15 people for an atmospheric tour of the collection. Incidentally, 15 is our preferred number as larger groups become too difficult to move around and as we like to keep the atmosphere relatively casual and be able to answer everybody’s questions. This photo shows a bit of context:

Photo event at the Royal Cast Collection
Photo event at the Royal Cast Collection

A couple of shots from the talented IG’ers:

Went on a cosy walk on SMK Cast Museum today – thanks for inviting us :) // A photo posted by Bobby Anwar (@bobbyanwar) on

Golden light

A photo posted by Oliver TCB (@olivertcb) on

Night at the museum… Thank You @igersdenmark for a fun night at The Royal Cast Collection. A photo posted by Abdellah Ihadian (@mr_babdellahn) on

For our next event we went anti-empty and invited photo folks for our grand opening of the “What’s Happening?” exhibition. The idea was to show the museum full of life and activity, which is a nice idea, but not necessarily one the IG’ers went for as you can see here:

Museum | kunst

A photo posted by Oliver TCB (@olivertcb) on

The | bag – thanks for the invite @smkmuseum ✌️ #smkmuseum #communityfirst #copenhagen

A photo posted by LittlemyCPH (@littlemycph) on

A photo posted by Oliver TCB (@olivertcb) on

But okay, you can’t always have (or plan) them all.

It’s been great seeing the museum through so many talented people’s lenses. We look forward to new adventures in square format in the time to come.

If you have any questions about our initiatives, don’t hesitate to ask.

Feel free to join our SMK photo events Facebook group.

For details, metrics and other wonderful things, see my Museums and the Web paper on The Me/Us/Them Model.

The new museum conversation is not about you

We used to fret over directing traffic to our own websites. Then we expanded our perspective to include all our channels. It’s time for the next leap: The one where we forget ourselves entirely.

It comes up all the time. The disparagement that museums (and indeed other institutions) are not fulfilling the true potential of social media. The idea – if you’ll allow me to paraphrase – that social media should usher in a golden age of democratic equality where museums are but partners in an enlightened, free-flowing dialogue. If you subscribe to this ideal, and then glance at the Facebook pages of your local museum you are likely to despair. What you see, very often, is at least partly marketing-like (“Oh no, look! They’re doing push communication!”) and dialogue is often sparse.

But here’s the thing: DON’T look too closely at those official feeds. Doing so is akin to your reporting in 2005, the time where your main measure of online success was traffic to your own website. Thankfully, we (mostly) managed to challenge this too-simple metric to include all our channels. Of course, we said, it does not make sense to think of the website as the ultimate destination when people can be inspired on Twitter, sign up for events on Facebook etc. It was the right argument at the right time. But now we need a similar shift in our priorities, our thinking, and our reporting.

Most of our social media strategies specify conversation and community-building as main goals of our efforts. In my view many museums are doing really important work to stimulate these conversations, but we don’t necessarily see it on our “owned” profiles. We see it as the result of making high-quality content easily available, of changing photo policies and of inspiring guests to talk and share. Museums provide building blocks and frameworks for conversation but these conversations largely take place in whatever contexts people find relevant.

In other words: We need another KPI revolution. We need to further minimize the us/them distinction. I like to say that “the strategy isn’t us, it’s them” but let me be clear: It’s really both. Museums can play an important role in stimulating conversations around their subject matter – a job that requires time and effort – but to measure that success in likes, engagement and mentions to some degree misses the point. We need to measure how our materials are used and the far more indirect ways in which we stimulate interest.

Is this broader conversation much harder to measure? Yes. Are our reporting tools tailored to this broader perspective? No. But if we allow ourselves to be guided by our tools we’d still only be comparing unique visits via Google Analytics.

On an endnote: To the skeptic this may sound like a cop-out. Like changing our reporting just because we couldn’t fulfil our goals. But it really isn’t. The goals remain (largely) the same, the effort required is equal or greater, and it’s only our old-fashioned – and rather self-centred – KPIs that need to change. And once we do, it’s very likely that we’ll stumble on all those conversations that some people vainly want to only appear in the official Facebook feed.

Frygt og lede i casual games-branchen

Små mobilspil med firkantede brikker. Lyder det ukontroversielt? Det er det langtfra, for hvornår er et brikspil et brikspil og hvornår er det nedrig udnyttelse af andres ideer?

I mobilspillet 2048 skal man skubbe til små brikker på en 4×4-felts spilleplade. Kombineres to brikker med det samme samme tal bliver de til én brik med de to oprindelige brikkers sum. Altså: Hvis du kombinerer to 4-taller, får du ét 8-tal. Når du til 2048 på én brik, har du vundet. Denne simple opskrift har taget App Store med storm og ude i verden kombinerer ganske almindelige mennesker lige nu talbrikker som var der ingen dag i morgen.

Og alle er glade, for det er da herligt uhøjtidelig hjernegymnastik.

Nej. Alle er ikke glade.

Jeg antydede en vis fascination af spillet på Twitter og fik denne kommentar:

Når 2048 er så frygteligt ukorrekt et tidsfordriv er det fordi spillet er en kynisk kopi af indie-spillet Threes, et værk skabt med hjerteblod og kreative kraftanstregelser. Det er i hvert fald én måde at se verden på. Men lad os lige starte med fakta:

Threes er udviklet af Asher Vollmer, Greg Wohlwend og Jimmy Hinson og publiceret på Apples App Store i januar 2014.

Threes. Tile-matching på en 4×4-spilleplade.

Her fik spillet dog ikke lov at brillere længe før en opportunistisk kinesisk herre uploadede det selverklæret Threes-inspirerede 1024.

1024. Tile-matching på en 4x4-spilleplade.
1024. Tile-matching på en 4×4-spilleplade.

1024 inspirerede meget hurtigt den 19-årige italienske udvikler Gabriele Cirulli til at fremstille sin egen variant: 2048.

2048. Tile-matching på en 4x4-spilleplade.
2048. Tile-matching på en 4×4-spilleplade.

Cirulli opfattede sit spil som en slags kopi af 1024 og forsøgte ikke at tjene penge på produktet. Tværtimod offentliggjorde han koden under en open source-licens, så alle kunne tilpasse 2048 efter deres hjertes lyst. Horder af udviklere gik direkte på opgaven, herunder lidet idealistiske folk hos firmaet Ketchapp, som publicerede en version på App Store med reklamer. Denne version blev voldsomt populær og Ketchapp har siden kunnet skumme fløden.

Det lyder jo unægteligt tyvagtigt og uretfærdigt og jeg kan sagtens forstå folk, som opfatter forløbet sådan. En del af mig er enig.

En anden del af mig tænker dog, at historien er lidt mere speget. For det første er Threes og 1024/2048 ikke identiske i deres grundmekanik. I Threes kan brikker kun kombineres ved at “slå” dem imod spillepladens kanter, ligesom der er adskillige begrænsninger på hvordan brikker kan kombineres. Denne kompleksitet er fraværende i 1024 og 2048, der bruger samme regelsæt (om end temmelig forskellig æstetik).

Men er det ikke bare detaljer? Måske. Men så bliver vi nødt til at snakke mere generelt om tile matching-spil (som spilforsker Jesper Juul har nærstuderet i denne artikel).

For vi har jo prøvet det før. F.eks. da du som barn prøvede at få fire på stribe i det populære plastikspil.

4 på stribe. Tile-matching på en 7x6-spilleplade.
4 på stribe. Tile-matching på en 7×6-spilleplade.

Eller dengang vi opgav nattesøvnen for at skabe fulde rækker af firkantede felter, der faldt ned fra oven (som i 4 på stribe) i Tetris. I øvrigt et spil, hvor jeg vist aldrig har spillet den helt originale russiske version og derfor allerede dengang var et dårligt menneske.

Tetris. Tile-matching på 11×20-spilleplade.

Det er dog kombinationen af nærliggende brikker som har vist sig mest populær. Tænk på Bejeweled, Candy Crush Saga og utallige børnespil (f.eks. Disney’s Frozen Free Fall).

Okay, Threes er ikke verdens første tile-mathing-spil, men dynamikken med kombinationen af talbrikker er da ny, ikke? Tjo..

Ideen med at kombinere  brikker for at få en “større” er gammelkeldt. Det var f.eks. et af principperne i iPhone-spillet Critter Crunch, hvor man fodrer små dyr til større dyr for at skabe endnu større dyr.

Critter Crunch.

Brikker med tal er jo heller ikke et ukendt fænomen. Tænk på gode gamle Minesweeper, hvor man skulle sandsynlighedsregne sig frem til hvor banens farlige felter var gemt.

Minesweeper. Tile-matching med tal.

Og så var der (absolut glimrende) Drop7, hvor brikker med tal svarende til antallet af brikker i rækken forsvandt.

Drop7. Tile-matching med tal på en 7×7-spilleplade.

Og så videre, og så videre.

Spørgsmålet er selvfølgelig: Når Tile-matching-spil er så udbredte og findes i så mange varianter, er det så rimeligt at hævde at et spil går over stregen og skamløst har kopieret et andet? Jeg vil lade det stå åbent. Efter min mening er 1024 en forbedring af Threes spilmekanik. Threes er en usleben diamant, men fandt ikke helt ind til kernen af den afhængighedsskabende, simple mekanik. 2048 er til gengæld en forbedring af 1024. Sidstnævntes æstetik er irriterende barnlig, mens 2048 går zen og renser spilverdenen for alt uvedkommende. Det er en slags Apple-destillation af spilprincipperne (når man ser bort fra de kyniske reklamer i gratis-versionen).

Når man forarges i dette tilfælde er det fordi den lille, hårdtarbejdende indie-udviklers arbejde bliver lynkopieret af grådige konkurrenter. Det er ikke rart at tænke på, og man kunne ønske sig en form for tvangsdeling af indtægterne. Men i dette tilfælde er kopierne betydeligt bedre end originalen og vi befinder os i et område hvor spilprincipperne er basale og udbredte i mange varianter at det er vanskeligt at afgøre hvor den originale ide begynder og slutter.

Se også:



Favoritbilleder fra 2013

Kameraerne  blev luftet lidt i 2013. Her er nogle personlige favoritter fra Flickr-strømmen.

Solopgang maler Rådhuspladsen rød en klar og kølig decembermorgen.
Khao Sok National Park
Højt bjerg, små mennesker – i Thailands Khao Sok-nationalpark.
Når den første tand ryger står den på stolthed i flere lag.
Den totale symmetri – solen skinner igennem Pantheon-kirkens kuppel i Rom.
Mere Pantheon. Der var noget med de her duer og hvor ligeglade de virkede med hele den antikke alvor. Jeg fangede det ikke helt, men det her blev ok.
Tietgen-kollegiet er rundt og stråler som guld. En fotogen combo.
Knald på farverne en sommeraften ved Cafe Metropolitain, Frederiksberg.
Livvagt ved Rørvig Strand. Ret ok job.
Lyset kan virke helt kunstigt en aften på sådan en smålandsk landevej.
Døtre på smålandsk stengærde.
Boblevand og vilde farver.
Makro, vind og en masse reflekser. Så er det næsten ok at motivet er blomster.
Nogle gange er lyset bare godt. Som her ved Peblingesøen i København.
Alvorssnak med cava i Torvehallerne, København.
Næsten abstrakt vinkel på Radiohusets facade, Frederiksberg.
Den Blå Planet på Amager. Fint sted.
Forladt karrusel lige ved lukketid.
Vanvittigt modlys og ukontrollabel glorieeffekt FTW.
Kongens Nytorv en tidlig morgen hvor kun duer og renovationsminiber er stået op.